# ParticipatingMedium Marks its entity as a volume of fog, smoke, dust or steam — something light has to travel through. What is behind the volume comes out dimmer than what is beside it, and overlapping volumes stack. ```luau local id = entity.spawn("fog_bank").id entity(id).component.add("ParticipatingMedium", { sizeX = 30, sizeY = 8, sizeZ = 30, density = 0.15, }) ``` ## Fields | Field | Meaning | |---|---| | `shape` | `"box"` (default) or `"sphere"` | | `density` | Extinction per world unit. `0` releases the medium. | | `sizeX` / `sizeY` / `sizeZ` | Half-extents, for a box | | `radius` | Radius, for a sphere | `density` is the physical knob: a ray crossing `d` world units of a medium of density `k` leaves with `exp(-k * d)` of the light it entered with. A 20-unit bank at `0.15` therefore passes about 5% of the light straight through it. The entity's position places the volume; move the entity and the shadow moves with it. ## Notes - The light-space map is sized to the bounding sphere of everything registered. A small bank sharing a scene with a very large one lands on few texels — set `extent` and `centre` through `mediaTransmittance.set` to bias the map toward the region you care about. - Attenuation applies to surfaces. The sky keeps its own colour. - The attenuated light is the sun's: `sunDirection` sets which way the map's rays travel. See `mediaTransmittance` for the scene-wide settings and the direct API.

# Volume Per-entity attachment that renders a 3D voxel texture (cloud, smoke, nebula, imported VDB grid) into the scene as a volumetric body. The component does NOT own the volume's lifetime — producers (the `volume` module, a custom compute fill, or `compute.writeTexture3D`) build the texture and pass its name in `volumeName`. The entity's world transform defines the AABB the raymarcher samples. `volumeName` is resolved through `Volume.get`, so a volume the `volume` module built is rendered as it stands — at its own dimensions, from its own density channel, through its occupancy grid, and out of its brick pool if it is paged. A name no volume answers to is taken as a bare 3D texture some other system owns, and `occupancyName` / `isRgba` describe it. ## Usage ```luau local Volume = require("@builtin::systems.volumetrics.volume") local cloud = Volume.noise({ name = "cloud", size = 64, scale = 4 }) cloud:buildOccupancy({ brickSize = 8 }) local e = entity.spawn("cloud_entity") e.localScale.set(8, 4, 8) e.component.add("Volume", { volumeName = "cloud", occupancyName = cloud.occupancyName, density = 6.0, }) ``` A paged volume costs what it holds and is named the same way: ```luau local cloud = Volume.noise({ name = "cloud", size = 128, threshold = 0.6 }) cloud:buildSparse({ brickSize = 8, threshold = 0.01 }) cloud:releaseDense() -- the brick pool is now the only copy local e = entity.spawn("cloud_entity") entity(e.id).localScale = { 8, 4, 8 } e.component.add("Volume", { volumeName = "cloud", density = 6.0 }) ``` ## Public fields | Field | Type | Sync | Default | Meaning | |---|---|---|---|---| | `volumeName` | string | Sync | `""` | Required. Name of the volume, or of a bare 3D texture, to render. | | `occupancyName` | string | Sync | `""` | R8 occupancy texture for a bare texture's empty-space skipping. Empty = no occupancy. A volume the module built carries its own, and a name set against one is reported as left aside. | | `density` | number | Sync | `1.0` | Density gain multiplier. | | `absorption` | number | Sync | `1.0` | Beer-Lambert absorption coefficient. | | `emission` | number | Sync | `1.0` | Emission gain. | | `color` | color | Sync | `{1,1,1,0}` | RGB tint + density floor in alpha. | | `stepSize` | number | NoSync | `0` | World units per ray step. `0` = auto-derive from volume extent. | | `maxSteps` | number | NoSync | `256` | Ray-march step cap. | | `target` | string | Sync | `"scene_volume_rt"` | Output render target name (composite key). | | `targetWidth` | number | NoSync | `1280` | Output width (px). | | `targetHeight` | number | NoSync | `720` | Output height (px). | | `useOccupancy` | boolean | NoSync | `true` | Toggle empty-space skipping. | | `isRgba` | boolean | Sync | `true` | `false` reads density from the red channel. Left `true`, a volume the module built is read from the channel it was built with, and a bare texture as RGBA. | All public fields use the `Field.<kind>(default, Sync|NoSync)` schema (PR #3019). Visual params (look, tint, density) propagate Sync; quality knobs (stepSize, maxSteps, target resolution, useOccupancy) are NoSync — each client picks its own perf/quality point. ## Lifecycle - **`awake`** — no-op. The volume texture is owned by the producer and looked up lazily on the first `update`. - **`update(dt)`** — resolves `volumeName` through `Volume.get`, falling back to a wrapper over a bare texture of that name, and dispatches the compute raymarcher into `target`. The resolution runs every frame, so a volume built, paged or destroyed after this component attached is picked up on the next one. - **`onPropertyChanged(key, value, old)`** — drops the resolved volume when `volumeName`, `occupancyName`, or `isRgba` changes, forcing a re-resolve on the next frame. ## Related - `volume` module — texture builders and name lookup (`Volume.noise`, `Volume.get`, `Volume.wrap`, `:buildSparse`, `:render`). - `vdb_to_zvol.importer` — VDB → dense voxel pipeline that produces textures consumable by this component. - `volumeDemos` / `volumetricSky` packages — worked examples.

# VolumeSequence Plays an open volume sequence through an entity's transform AABB. The sequence is opened by a producer and named here — the same split the `Volume` component makes between "what data to render" and "where it appears in the world". Each tick the component advances the playhead and raymarches the frame that is resident into the shared `scene_volume_rt` target, in both play and edit mode, so the run animates while it is being authored. ```luau local Seq = require("@builtin::systems.volumetrics.volumeSequence") Seq.open({ name = "plume", frames = paths, resident = 3, fps = 24 }) local id = entity.spawn("plume_entity").id entity(id).localScale = { 8, 8, 8 } entity(id).component.add("VolumeSequence", { sequenceName = "plume", density = 6.0 }) ``` `fps` and `playing` are pushed into the sequence when they change, so a script calling `seq:setFps()` / `seq:pause()` between two ticks is left standing until the authored value itself moves. The component does not own the sequence: despawning the entity leaves the ring allocated, and `volumeSequence.close(name)` is what frees it. ## Sharing The playhead, the rate and the ring belong to the sequence, so several entities naming one run play the same frame from the same three textures: the playhead advances once per engine frame however many components tick it, and the rate is whichever component's `fps` moved last. Give each entity its own run when they should play independently. `target` is a render target, and a raymarch writes the whole of it — two volumes visible at once (a sequence beside a `Volume`, or two sequences) need a `target` name each, or the later write replaces the earlier one and only one of them appears. Composite each target the way any volume render is composited — see `volume.compositeShaderSource()`.

# VolumetricLighting The air between the camera and the scene, as a component. Add it to an entity and light travelling through that air scatters toward the viewer, so a lamp becomes a glow rather than an invisible point and the sun cuts shafts where something crosses it. ```lua local id = entity.spawn("air").id entity(id).component.add("VolumetricLighting", { density = 0.06, anisotropy = 0.7, }) ``` Every field is live — editing one in the inspector changes the air on the next frame, and the values persist with the scene. Setting `density` to 0 empties the air and releases the passes while keeping the rest authored; removing the last one in the scene clears it. | field | effect | |---|---| | `density` | How much medium there is per world unit at the reference altitude. 0 is off. | | `heightFalloff` / `heightRef` | How fast density thins with altitude, and the altitude `density` is quoted at. 0 falloff fills the scene evenly. | | `steps` | Samples along each view ray. More resolves a beam's edge more cleanly. | | `colorR/G/B` | What the medium scatters. Warm dust and cold mist differ here. | | `anisotropy` | How forward-biased the scattering is, in (-1, 1). Positive values throw light onward, so haze blooms around a source you look toward. | | `sunIntensity` / `sunColorR/G/B` | The sun's contribution to the medium. | | `lightScale` | The scene lights' contribution to the medium. | | `shadowSteps` / `shadowDistance` | Samples taken back toward the sun to decide whether a point in the air is lit, and how far they reach. **0 leaves the sun's contribution unshadowed**, so the haze glows evenly and no shafts form. | | `maxDistance` | How far along the view ray the march runs. A ray that escapes to sky stops here. | The lights are the scene's own, each scattering on the terms its kind emits on: a point fills the air around it, a spot fills its cone, a rect reaches the air on the side its face emits along, and a parallel light lays an even glow across the whole medium. One added, moved, re-aimed or recoloured at runtime changes the glow it throws with nothing to keep in step here. One body of air fills the scene, so one of these per scene is enough — the last one to push wins. The work runs in `lightScattering.renderFeature`; the settings live in `@builtin::systems.volumetrics.volumetricLighting`, which is also how to drive them from a script.

# VolumetricSky Singleton component that drives the procedural cloud layer. Attach to one entity (usually a sky placeholder), and the component owns the cloud volume + render target lifecycle, bakes the noise field when what it is baked from changes, and dispatches the compute raymarcher every frame into `scene_volumetric_sky`. A composite (post-process or material) blends `scene_volumetric_sky` over the scene — see `demo/sky.scene/` for the full wiring. ## Usage ```luau -- Defaults (the v4-tuned cumulus look — `Sky.defaultPreset()` hands out a -- copy of the same table this field defaults to) local sky = entity.spawn("sky") sky.component.add("VolumetricSky", {}) -- Or apply a captured preset: preset.apply("cumulus_default", { entity = sky.id, component = "VolumetricSky" }) -- Inline tweak. `component.add` writes the `preset` field as it is given, -- so a table handed here carries the whole configuration: local p = require("@builtin::systems.volumetrics.volumetricSky").defaultPreset() p.noiseScale, p.octaves, p.coverage = 9.0, 5, 0.20 p.densityGain = 0.04 -- v5 σ_t sky.component.add("VolumetricSky", { shaderVariant = "v5", preset = p }) ``` `volumetricSky.module`'s `M.spawn` / `M.configure` take the same parameters by name, folding each into the current configuration and leaving the rest where they stand: ```luau local Sky = require("@builtin::systems.volumetrics.volumetricSky") Sky.spawn({ shaderVariant = "v5", noiseScale = 9.0, coverage = 0.20 }) ``` ## Schema (Field<T> / Sync / NoSync) The 28 shaping / weather / lighting params travel as ONE Sync field — `preset = Field.table(DEFAULT_PRESET, Sync)` — so replicating a sky change is **one mutation, not 28**, and copying a sky between scenes is **copying one field, not 28**. Per-client quality knobs sit outside the preset as NoSync fields so each client picks its own tier without disturbing peers. `SkyPreset` and `DEFAULT_PRESET` are declared by `volumetricSky.module`, which this component reads its default from and which routes an option by that same key set — so the shape below and the module's `M.spawn` options are one declaration. | Field | Type | Sync | Default | Meaning | |---|---|---|---|---| | `preset` | `SkyPreset` table | Sync | `DEFAULT_PRESET` (v4 cumulus) | All shaping / weather / lighting (see below). | | `shaderVariant` | string | Sync | `"v4"` | `"v3" \| "v4" (default) \| "v5"`. | | `wind` | vec3 | Sync | `{0,0,0}` | Animated drift offset (advanced per frame from `preset.windSpeed`). | | `cloudSize` | number | NoSync | `256` | Volume W,D resolution. | | `cloudHeight` | number | NoSync | `128` | Volume H resolution. | | `stepSize` | number | NoSync | `4.0` | World units per ray step. | | `maxSteps` | number | NoSync | `224` | Ray-march step cap. | | `shadowSteps` | number | NoSync | `8` | Shadow ray step count. | | `alphaCutoff` | number | NoSync | `0.99` | Early-exit α threshold. | | `jitter` | number | NoSync | `1.0` | Ray-origin jitter strength. | | `taaBlendAlpha` | number | NoSync | `0.1` | TAA current-sample weight. | | `targetWidth` | number | NoSync | `0` | Render target width (`0` = viewport). | | `targetHeight` | number | NoSync | `0` | Render target height. | | `noiseRegenRate` | number | NoSync | `10` | Times per second wind drift re-bakes the cloud volume (`0` = every frame the wind moved). | ### Preset (`SkyPreset`) shape Noise / shaping: `noiseScale, octaves, coverage, densityGain, altitudeLow, altitudeHigh, anvilBias, detailScale, detailOctaves, detailStrength, detailThreshold, baseLowRemap, baseHighRemap, color, windSpeed, seed`. Volume placement: `horizExtent, cloudBaseAltitude, cloudTopAltitude`. Look (visual, not perf): `phaseG`. Sun / sky: `sunDir, sunColor, sunIntensity, sunDiscSize, horizonColor, zenithColor, groundColor, ambient`. v5-only: `cloudType, worleyBlend`. See `presets/cumulus_default.preset/preset.yaml` for the canonical full preset shape, and the `init.luau` `DEFAULT_PRESET` table for the runtime defaults. ## Where the frame time goes Two passes carry the cost, and different fields reach each: - **The bake** (`cloud_noise_*`) fills every voxel of the `cloudSize × cloudHeight × cloudSize` volume with layered noise. It is the most expensive pass the layer has (around 10 ms at 256×128×256 on a desktop GPU), and it runs only when the volume's inputs move: at once for a change to `preset`, `cloudSize`, `cloudHeight`, `shaderVariant` or the render device, once the program it compiles is resident for an edit to the noise shader's own source, and at most `noiseRegenRate` times per second for wind drift. A sky whose `preset.windSpeed` is zero bakes once. `cloudSize` / `cloudHeight` set what one bake costs; `noiseRegenRate` and `windSpeed` set how often it is paid. - **The raymarch** (`cloud_raymarch_*`) runs for every view of every frame. `stepSize`, `maxSteps`, `shadowSteps`, `alphaCutoff` and `targetWidth` / `targetHeight` set its cost. `profiler.gpuFrame()` reports both as `compute.<shader>` spans; a span's `frames` count is how many frames of its window the pass ran in, and `volumetricSky.volumeBakes()` counts the bakes themselves. ## Field-mutation gotcha (and why `update()` reassigns the whole vec3) Under the new Field<T> / Sync / NoSync schema, `public` is not a plain Lua table — it's wrapped in a proxy whose `__newindex` metamethod performs type-checks, marks the field dirty, and queues Sync replication. Writing `public.wind = {x,y,z}` triggers all four steps. Writing `public.wind[1] = x` reads `public.wind` (getter), mutates the returned table directly, and **bypasses the setter** — no Sync replication, no dirty flag, no `onPropertyChanged`. Always assign the whole vec3 / preset / color table; never index-mutate. The same caveat applies to `public.preset`: `public.preset.coverage = 0.2` won't replicate. To change one preset field, copy the table, mutate, reassign: ```luau local p = public.preset local next = {} for k, v in pairs(p) do next[k] = v end next.coverage = 0.2 public.preset = next ``` ## Lifecycle - **`awake`**: starts the sky's clock and creates the render feature that runs the layer's passes. - **`update(dt)`** — advances `wind` (reading `preset.windSpeed`), hands the noise inputs to `volumetricSky.fillClouds`, which bakes the volume when they moved (see "Where the frame time goes"), then queues the raymarch dispatch for the frame's deferred phase. - **`onDespawn`** — releases the volume, the render target and the composite effect via `M.destroy`. ### Why the raymarch is dispatched from `task.defer` The raymarch produces ONE full-screen layer for ONE view, and the frame composites it over everything that frame drew — so it has to be rendered from the view the frame ends up drawn with. Which view that is settles only once the frame's script work is finished: a cutscene walking the camera from a task loop, a camera controller, a component ordered after this one, an editor tool — each writes the camera after `update` has run, while the renderer reads the camera once all of them are done. Taking the camera snapshot in the deferred phase reads what every one of them has already written, and still queues the dispatch ahead of the frame's render. Snapshotting in `update` instead renders the layer for whatever the camera was midway through the frame; on a hard cut that is a different place entirely, and the frame the cut lands on composites a cloud layer belonging to the shot it just left. A caller driving `volumetricSky.dispatchRaymarch` directly owes the same discipline. ## Related - `volumetricSky.module` — orchestration backend. - `presets/cumulus_default.preset` — canonical full preset; copy to author named weather (storm, overcast, dawn, …). - `Volume.component` — per-entity flow for non-cloud volumetric content. - `demo/sky.scene` — worked example.

# godrays_fill Fills a volume with shafted light. One invocation per voxel. ## Bindings | name | kind | what it is | |------|------|------------| | `vol` | `storage3d` (rgba16f) | the volume written — RGB emission, A density | | `U` | `buffer` (read, f32) | the demo's fill parameters |

# nebula_fill Fills a volume with a spiral nebula: RGB carries emission, A carries density. One invocation per voxel at workgroup size 4x4x4. ## Bindings | name | kind | what it is | |------|------|------------| | `vol` | `storage3d` (rgba16f) | the volume written, read-write so a pass can accumulate | | `U` | `buffer` (read, f32) | the demo's fill parameters, in the order the shader reads them | `U` holds: elapsed time, spiral frequency, arm thickness, core brightness, edge brightness, dust density, noise strength.

# sdf_solids_fill Fills a volume with signed-distance solids. One invocation per voxel. ## Bindings | name | kind | what it is | |------|------|------------| | `vol` | `storage3d` (rgba16f) | the volume written — RGB emission, A density | | `U` | `buffer` (read, f32) | the demo's fill parameters |

# smoke_plume_fill Fills a volume with a rising smoke plume. One invocation per voxel. ## Bindings | name | kind | what it is | |------|------|------------| | `vol` | `storage3d` (rgba16f) | the volume written — RGB emission, A density | | `U` | `buffer` (read, f32) | the demo's fill parameters |

# zvol_viewer_fill Leaves the volume as it is. The ZVOL viewer uploads voxels straight into the volume's 3D texture from the CPU, so the fill stage this shader occupies has nothing to compute — it reads one value from each binding so both stay live. ## Bindings | name | kind | what it is | |------|------|------------| | `vol` | `storage3d` (rgba16f) | the volume the viewer writes into from the CPU | | `U` | `buffer` (read, f32) | the fill parameters the spawn API supplies |

# emission_raymarch Marches a camera ray through an emissive volume and writes the result to a 2D target. Emission accumulates along the ray and density attenuates what is behind it, so a volume reads as light rather than as a surface. ## Bindings | name | kind | what it is | |------|------|------------| | `vol` | `texture3d` | the volume marched — RGB emission, A density | | `smp` | `sampler` | how `vol` is filtered along the ray | | `out_tex` | `storage2d` (rgba16f) | the image the march writes | | `U` | `buffer` (read, f32) | camera, volume placement and march settings | `U` carries the inverse view-projection, the volume's world transform, the step count, density and emission gains, and the target's inverse size.

# VDB → ZVOL importer Claims `.vdb` (OpenVDB) writes and produces a sibling `.zvol` — the engine-native dense RGBA8 volumetric format consumed by the volumetrics system. This is a standard `importer` asset: `init.luau` returns a `{ canImport, import }` module that the importer dispatcher (`@builtin::assetTypes.importer`) loads lazily and runs only when a matching file is written. It does **not** execute anything at load time. ## Pipeline 1. **Real parse** — if the `openvdb` plugin in the builtin library (`@builtin::plugins.openvdb`) is loaded, the VDB B-tree is parsed into a dense RGBA8 voxel buffer and wrapped in a 64-byte ZVOL header. 2. **Procedural fallback** — without the plugin, a deterministic noise-modulated torus is synthesised (FNV-1a hash of the first 256 source bytes seeds it) so the importer + demo wiring stays functional. The `produced_by` field in the result distinguishes the two paths. ## ZVOL header (64 bytes) | offset | size | field | |--------|------|-------| | 0 | 4 | magic `ZVOL` | | 4 | 4 | u32 version (1) | | 8 | 4 | u32 width | | 12 | 4 | u32 height | | 16 | 4 | u32 depth | | 20 | 4 | u32 format (1 = rgba8) | | 24 | 12 | f32[3] aabb_min | | 36 | 12 | f32[3] aabb_max | | 48 | 16 | reserved | | 64 | … | voxel data, `w*h*d*4` bytes, XY-major Z-slowest |

# lightScattering The render passes behind the `VolumetricLighting` component and the `volumetricLighting` module. Two compute passes after lighting. **`vol_scatter`** marches each view ray at half resolution, from the camera to the surface the ray hits — or to the medium's reach when it escapes to sky, since an unbounded march would integrate the whole horizon into every sky pixel. At each step it evaluates the medium's density, gathers the light arriving there, and accumulates the fraction sent toward the camera, attenuating by everything already crossed. A surface-only renderer draws the pool of light a lamp throws on the floor but not the beam that carried it there, because nothing between the lamp and the floor is ever shaded. Marching the ray shades that space. The lights are the scene's own. `@scene.lights` is the same buffer the lit passes read, so a light added, moved or recoloured at runtime changes the beam it throws with no bookkeeping here. It holds **twelve vec4s per light** — the light, then its punctual shadow record — read as raw words so no struct declaration has to mirror the engine's layout. A spot's cone gate is evaluated at the sample itself rather than at a surface, which is what makes the cone visible as a volume instead of only where it lands. Attenuation matches the lit passes' windowed falloff, so a beam and the pool of light it lands in agree about how far the source reaches. Area lights are skipped — their contribution to a medium is an integral over the rect rather than a single direction. The sun's contribution is gated by a short march back toward it, tested against the depth buffer. That is what turns uniform haze into shafts: the medium is lit only where the path back to the sun is clear, so an occluder carves a dark volume out of the beam rather than only shading the surface behind it. Because the test reads the depth buffer, an occluder outside the view frustum does not cast into the medium. Scattering follows Henyey-Greenstein, so the medium brightens toward a source you are looking at and stays dim when it is behind you — the difference between haze that reads as air and haze that reads as a grey wash. Each ray's first sample is offset by a per-pixel fraction of a step, trading the banding a fixed step pattern produces for noise. **`vol_composite`** upsamples that result and composites it. The taps are weighted by how closely their depth matches the full-resolution pixel's, which keeps a beam from bleeding across a silhouette — the halo a plain bilinear upsample leaves behind. The camera comes from `@frame.camera`, the camera of the render target being drawn, and the forward matrix is derived by inverting it in the shader rather than passed in from script — so an offscreen capture or an RTT camera marches through its own matrix at its own resolution. Parameters arrive through the `vol_params` buffer, packed by the `volumetricLighting` module.

# mediaShadow The shadow a fog bank casts on what is behind it. Two compute passes run after lighting: | Order | Pass | Work | |---|---|---| | 36 | `media_tau_map` | One ray per texel of a light-space map; accumulate optical depth | | 37 | `media_shadow_apply` | Attenuate each surface by the density in front of it | | 38 | composite | Blit the result | ## Why not a shadow map A depth-based shadow map answers one question per texel: is this point nearer the light than the first thing the light hit? Density has no first thing. What a medium does to light is continuous — a ray crossing `d` units of extinction `k` keeps `exp(-k*d)` of its energy — so what a receiver needs is the optical depth between it and the light, not a depth to compare against. The first pass therefore stores optical depth rather than distance, at four points down each ray. Two consequences fall out of that choice: - **Overlapping media stack.** Optical depth is additive, so two volumes that intersect contribute to the same running total with nothing to resolve. - **Geometry inside a volume is lit correctly.** A receiver looks up its own position along the ray and reads the depth accumulated *up to that point*, so a post standing in the middle of a fog bank is dimmed by the half in front of it rather than by the whole column. ## Inputs `@scene.color`, `@scene.depth`, `@frame.camera`, and the `media_params` / `media_volumes` buffers packed by `@builtin::systems.volumetrics.mediaTransmittance`. Sky pixels are passed through untouched.

# mediaTransmittance Light attenuation through participating media, in a form both media and surfaces can sample. A shadow map stores one depth per texel and answers a yes/no question, so nothing made of density can cast with it. This module owns the other representation: a light-space map holding the **optical depth** accumulated along each light ray, recorded at four points down its length. Optical depth is additive. Every medium a ray crosses contributes to the same texel, so overlapping volumes stack with nothing to resolve between them. And because four points are stored rather than one total, a receiver standing halfway through a fog bank is attenuated by the half in front of it, not by the whole column. ## Direct use ```luau local Media = require("@builtin::systems.volumetrics.mediaTransmittance") Media.set({ sunDirection = { 0.55, 0.55, 0.55 }, steps = 32 }) Media.setMedium("bank", { position = { 0, 8, 0 }, size = { 20, 6, 20 }, density = 0.15, }) ``` Most scenes attach `ParticipatingMedium` to an entity instead and let it drive this module. ## API | Call | Effect | |---|---| | `set(opts)` | Scene-wide settings; omitted fields keep their value | | `setMedium(key, opts)` | Register or move a medium under a stable key | | `removeMedium(key)` | Drop one medium | | `get()` | A copy of the settings in force | | `clear()` | Drop everything and release the passes | | `active()` | Whether the passes are running | ### Settings | Field | Meaning | |---|---| | `sunDirection` | Points **toward** the sun; light travels the other way | | `strength` | How much of the computed attenuation is applied, `[0, 1]` | | `steps` | Samples along each light ray, `[4, 128]` | | `extent` | Half-width of the map in world units; `0` sizes it to the media | | `centre` | Point the map is centred on; omit to centre it on the media | ### Media | Field | Meaning | |---|---| | `position` | World-space centre | | `shape` | `"box"` (default) or `"sphere"` | | `size` | Half-extents, for a box | | `radius` | Radius, for a sphere | | `density` | Extinction per world unit | Up to 32 media are packed into the records buffer. ## Cost and limits The map is 512×512 and is rebuilt each frame; `steps` sets the march length. With nothing registered the passes are released entirely, so a scene with no media pays nothing. The map is sized to the bounding sphere of everything registered, so a small bank sharing a scene with a much larger one lands on few texels — set `extent` and `centre` explicitly to bias the map toward the region that matters. Attenuation applies to surfaces. The sky keeps its own colour.

# media_common The media-shadow passes' shared ground: the volume record layout, the transmittance accumulation along a ray, and the cap on how many volumes one frame carries (`MAX_VOLUMES` = 32). That cap sizes a fixed array, so it is a constant here rather than a param.

# media_shadow_apply Applies the media's transmittance to the lit scene. ## Bindings | name | kind | |------|------| | `source_map` | `texture2d` | | `scene_depth` | `texture_depth` | | `tau_map` | `texture2d` | | `frame_cam` | `buffer` | | `media_params` | `buffer` | | `media_out` | `storage2d (rgba16f)` |

# media_tau_map Builds the transmittance map through the media, so a later pass reads how much light survives to a point instead of marching for it. ## Includes - `media_common` ## Bindings | name | kind | |------|------| | `media_params` | `buffer` | | `media_volumes` | `buffer` | | `media_tau_out` | `storage2d (rgba16f)` |

# cumulus_default Default weather preset for `VolumetricSky.component`. Captures the v4-tuned cumulus look: visible discrete clouds, warm-afternoon sun, moderate coverage. ## Use ```luau local sky = entity.spawn("sky") sky.component.add("VolumetricSky", {}) preset.apply("cumulus_default", { entity = sky.id, component = "VolumetricSky" }) ``` The values set the component's `preset` Field.table as a unit — one Sync replication carries every shaping/lighting parameter to peers. Tweak by reading the values and changing them before the component is added: ```luau local values = preset.values("cumulus_default") values.shaderVariant = "v5" -- swap raymarcher values.preset.coverage = 0.6 -- sparser sky sky.component.add("VolumetricSky", values) ``` ## Authoring more presets Capture a sky you have tuned with `preset.capture({ entity = sky.id, component = "VolumetricSky" }, "<name>")`, or drop another `<name>.preset/` folder next to this one (or anywhere under `/zero/source/`); the asset indexer registers it and `preset.values("<name>")` resolves it. Common alternatives: - `stratocumulus.preset` — broad overcast layer (`coverage ~= 0.15`, `cloudType = 0.5`, `altitudeHigh ~= 0.65`). - `clear.preset` — barely any cloud (`coverage ~= 0.7`, `densityGain ~= 0.2`). - `storm.preset` — dense black anvils (`coverage ~= 0.1`, `densityGain ~= 0.85`, `sunColor ~= [0.6, 0.55, 0.5]`). `preset` is one `Field.table`: applying a preset replaces it whole, so provide every field.

# scattering_common The light-scattering passes' shared ground: the phase function, the shadow lookup along a ray, and the froxel-to-world mapping. Included with `#include "@builtin::systems.volumetrics.scattering_common"`.

# cloud_noise Procedural volumetric-cloud noise fill (Schneider/Hillaire style). A low-frequency FBM gives the base blob silhouette, a high-frequency FBM erodes it near the surface for the fluffy cumulus look, plus altitude shaping (round bottom, full middle, anvil top). Writes the 3D cloud density texture. Bindings (see bindings.yaml): `clouds` (storage3d rgba16f — the cloud volume), `U` (read `array<f32>` uniforms). Usage: dispatch by identity with `compute.dispatchEx("@builtin::systems.volumetrics.shaders.cloud_noise", ...)` (storage texture requires resource binding, not positional buffers) — it resolves to the shader's stable guid and auto-compiles on first use (no setup).

# cloud_noise_v3 Volumetric-cloud noise fill (Schneider/Hillaire style), v3 variant. Low-frequency FBM base silhouette + high-frequency FBM surface erosion for the billowed cumulus look, with altitude shaping (round bottom, full middle, anvil top). Writes the 3D cloud density texture. Bindings (see bindings.yaml): `clouds` (storage3d rgba16f — the cloud volume), `U` (read `array<f32>` uniforms). Usage: dispatch by identity with `compute.dispatchEx("@builtin::systems.volumetrics.shaders.cloud_noise_v3", ...)` (storage texture requires resource binding, not positional buffers) — it resolves to the shader's stable guid and auto-compiles on first use (no setup).

# cloud_noise_v5 Volumetric-cloud noise generator, v5 (Nubis/Schneider style). Builds two foundational noises — Perlin-Worley FBM for the lumpy "cauliflower" base shape and Worley FBM for high-frequency detail erosion — under weather-map control. Writes the 3D cloud density texture. Bindings (see bindings.yaml): `clouds` (storage3d rgba16f — the cloud volume), `U` (read `array<f32>` uniforms). Usage: dispatch by identity with `compute.dispatchEx("@builtin::systems.volumetrics.shaders.cloud_noise_v5", ...)` (storage texture requires resource binding, not positional buffers) — it resolves to the shader's stable guid and auto-compiles on first use (no setup).

# cloud_raymarch Volumetric-sky cloud raymarcher with Frostbite-style lighting: Wrenninge/Hillaire multi-scatter approximation (3 octaves), a 5-sample cone-traced soft self-shadow, and combined Beer-Lambert + Powder so cloud edges glow toward the sun. Reads the cloud volume + occupancy textures and writes the lit raymarch result. Bindings (see bindings.yaml): `clouds` (texture3d), `occupancy` (texture3d), `cloud_smp` (sampler), `out_tex` (storage2d rgba16f), `U` (read `array<f32>` uniforms). Usage: dispatch by identity with `compute.dispatchEx("@builtin::systems.volumetrics.shaders.cloud_raymarch", ...)` (textures/sampler/storage texture require resource binding, not positional buffers) — it resolves to the shader's stable guid and auto-compiles on first use (no setup).

# cloud_raymarch_v3 Volumetric-sky cloud raymarcher (Frostbite-style lighting), v3 variant: Wrenninge/Hillaire multi-scatter (3 octaves), 5-sample cone-traced soft shadow, and combined Beer-Lambert + Powder edge glow. Reads the cloud volume + occupancy textures and writes the lit raymarch result. Bindings (see bindings.yaml): `clouds` (texture3d), `occupancy` (texture3d), `cloud_smp` (sampler), `out_tex` (storage2d rgba16f), `U` (read `array<f32>` uniforms). Usage: dispatch by identity with `compute.dispatchEx("@builtin::systems.volumetrics.shaders.cloud_raymarch_v3", ...)` (textures/sampler/storage texture require resource binding, not positional buffers) — it resolves to the shader's stable guid and auto-compiles on first use (no setup).

# cloud_raymarch_v4 Volumetric-sky cloud raymarcher, v4. Adds procedural detail noise computed in-shader on top of the large-scale texture silhouette (removes v3's voxel-grid blockiness) and uses multi-distance shadow sampling (4 geometrically spaced strides) instead of a cone trace. Reads the cloud volume + occupancy textures and writes the lit raymarch result. Bindings (see bindings.yaml): `clouds` (texture3d), `occupancy` (texture3d), `cloud_smp` (sampler), `out_tex` (storage2d rgba16f), `U` (read `array<f32>` uniforms). Usage: dispatch by identity with `compute.dispatchEx("@builtin::systems.volumetrics.shaders.cloud_raymarch_v4", ...)` (textures/sampler/storage texture require resource binding, not positional buffers) — it resolves to the shader's stable guid and auto-compiles on first use (no setup).

# cloud_raymarch_v5 Production volumetric-cloud raymarcher, v5 (Nubis/Hillaire style): Hillaire multi-scatter octaves (fixes black cloud cores), energy-conserving Beer-Powder-MS lit-edge glow, a 6-offset Wrenninge cone-trace shadow, and a dual-lobe Henyey-Greenstein phase. Reads the cloud volume + occupancy textures and writes the lit raymarch result. Bindings (see bindings.yaml): `clouds` (texture3d), `occupancy` (texture3d), `cloud_smp` (sampler), `out_tex` (storage2d rgba16f), `U` (read `array<f32>` uniforms). Usage: dispatch by identity with `compute.dispatchEx("@builtin::systems.volumetrics.shaders.cloud_raymarch_v5", ...)` (textures/sampler/storage texture require resource binding, not positional buffers) — it resolves to the shader's stable guid and auto-compiles on first use (no setup).

# cloud_taa Cloud TAA composite — temporal anti-aliasing for the raymarch output. Reprojects last frame's blended output through the previous view-projection, color-clamps the history to the current sample's neighborhood to suppress ghosting, and exponentially blends the two. Writes both the next-frame history slot and the externally-sampled `scene_volumetric_sky` texture the composite path reads. Reprojection runs through the world point named by uniform `[37]`, the NDC z of the range the cloud stands at along the view ray — `volumetricSky.reprojectionDepthNdc` computes it. The weight the history carries falls to zero across the last `REPROJECT_BORDER_TEXELS` of the reprojected frame, so a pixel whose history lies off the previous frame reaches the raymarch's own output down a gradient rather than across a line. Bindings (see bindings.yaml — slot order must match the Luau dispatch): `cur_tex` (texture2d), `hist_prev` (texture2d), `smp` (sampler), `out_hist` (storage2d rgba16f), `out_scene` (storage2d rgba16f), `U` (read `array<f32>` uniforms). Usage: dispatch by identity with `compute.dispatchEx("@builtin::systems.volumetrics.shaders.cloud_taa", ...)` (textures/sampler/storage textures require resource binding, not positional buffers) — it resolves to the shader's stable guid and auto-compiles on first use (no setup).

# volume_pack Fills the brick pool from the volume `volume_pages` has just paged. One invocation per pool voxel. The pool is a grid of blocks, one block per slot; a block holds a brick plus a one-voxel apron taken from the neighbouring bricks, so a trilinear sample inside a brick reads what the dense volume would have given it rather than whatever brick the pool stores next door. ## Bindings | # | name | what | |---|---|---| | 0 | `source` | the dense volume being paged (sampled 3D texture) | | 1 | `atlas` | the brick pool (`rgba16f` storage) | | 2 | `page_slots` | four words per slot, written by `volume_pages` | | 3 | `K` | `PackParams` — volume dimensions and brick edge | ## Why the pool is always `rgba16f` A pool is written as a storage texture and sampled with filtering. `rgba16f` is the format that is both: the one-channel formats cannot be written as storage textures, and `r32f` cannot be filtered without a device feature the engine does not require. So a scalar volume's density rides in the pool's red channel — which is where the raymarcher reads density from in density mode — and pays `rgba16f` per resident voxel. `Volume:sparseStats()` reports the dense and resident sizes side by side rather than assuming paging is the smaller of the two. The pass writes every voxel of the pool, including blocks whose slot was never handed out: those hold whatever brick their stale slot entry names, and the page table never points at them.

# volume_pages Decides which bricks of a volume are worth storing, and where each one lives. One invocation per brick. It scans the brick's voxels in the source volume, and a brick holding anything over `threshold` claims a slot from an atomic counter. The slot number lands in the page table (`slot + 1`, so 0 keeps meaning "empty"), and the brick's grid coordinates land in the slot list so `volume_pack` knows what to copy into that slot. Nothing is read back to the CPU and no brick waits on another, so the pass can run every frame — which is what makes the page table usable as the acceleration structure for a volume whose contents change while it renders. ## Bindings | # | name | what | |---|---|---| | 0 | `source` | the dense volume being paged (sampled 3D texture) | | 1 | `pages` | page table, one texel per brick (`rgba16f` storage) | | 2 | `page_alloc` | `[0]` slots handed out, `[1]` bricks that found the pool full | | 3 | `page_slots` | four words per slot: the brick's x, y, z, and 1 | | 4 | `P` | `PageParams` — volume dimensions, brick edge, threshold, channel, capacity | `page_alloc` must be zeroed before each run; the counter is what allocates, so a stale count hands out slots past the pool. ## The page value A resident brick's page value is `slot + 1`, split into two digits base 1024 across the texel's red and green channels. The page table is `rgba16f` because that is the format that is both storage-writable and filterable — the raymarcher binds it where a sampled 3D texture is expected — and a half float only carries integers exactly up to 2048, which one digit stays under. A brick holding nothing writes zero in both channels, so "is anything here" is the sum against zero, the same test the older single-channel occupancy grid answers. ## Overflow A brick that arrives after the pool is full is counted in `page_alloc[1]` and written as empty. The volume renders without it rather than sharing a slot. `page_alloc[0]` keeps counting past the capacity, so a caller that reads it back learns the capacity the volume actually needs.

# volume_raymarch Generic volume raymarcher. One invocation per output pixel: builds a world-space ray from the camera inverse view-projection, clips against the volume AABB, and marches front-to-back, using the occupancy grid to skip empty bricks. Output is RGBA16F (premultiplied RGB, accumulated alpha) for the caller to composite. A paged volume marches through the same shader: `volume` is then the brick pool and `occupancy` the page table, whose texel per brick carries the pool slot that brick lives in (0 for a brick holding nothing, which is what the empty-space skip already reads it as). `VolumeUniforms.paged` carries the flag and the volume's own dimensions, which the pool no longer describes. Both the skip and the pool lookup find a point's brick by the voxel it lands on divided by the brick edge — the way the page table and the occupancy grid were written. A volume that is not a whole number of bricks across has a grid wider than `dims / brick`, so dividing the normalized coordinate by the grid instead would name a different brick for the same point. When the dimensions are unknown (a wrapper was never told them) the grid times the brick edge stands in for them. Bindings (see bindings.yaml): `volume` (texture3d), `occupancy` (texture3d), `volume_smp` (sampler), `out_tex` (storage2d rgba16f), `U` (read buffer, a single `VolumeUniforms` struct). Usage: dispatch by identity with `compute.dispatchEx("@builtin::systems.volumetrics.shaders.volume_raymarch", ...)` (textures/samplers/storage texture require resource binding, not positional buffers) — it resolves to the shader's stable guid and auto-compiles on first use (no setup).

# vol_composite Composites the scattered light over the scene, depth-aware so geometry occludes it. ## Includes - `scattering_common` ## Bindings | name | kind | |------|------| | `scene_color` | `texture2d` | | `scene_depth` | `texture_depth` | | `scatter_map` | `texture2d` | | `vol_params` | `buffer` | | `vol_out` | `storage2d (rgba16f)` |

# vol_scatter Accumulates in-scattered light through the volume, one froxel at a time. ## Includes - `scattering_common` ## Bindings | name | kind | |------|------| | `scene_depth` | `texture_depth` | | `frame_cam` | `buffer` | | `lights` | `buffer` | | `vol_params` | `buffer` | | `scatter_out` | `storage2d (rgba16f)` |

# volume High-level Luau API for 3D textures and volumetric rendering. Composes the engine's generic `compute.*` primitives into volume-specific helpers — procedural builders (`sphere`, `box`, `noise`), a sparse NanoVDB-style occupancy grid for empty-space skipping, and a one-call raymarcher that composites a volume through any entity's transform. Pairs with the `Volume` component which attaches a built (or imported) texture to an entity. ## Exports - `M.create({name, width, height, depth, format?, storage?}) -> Volume` — allocate a fresh, zero-filled 3D texture and return a Volume wrapper. - `M.wrap({name, format, occupancyName?, occupancyBrick?}) -> Volume` — wrap an existing named texture without recreating it (used by the .zvol viewer). - `M.get(name) -> Volume?` — the live volume built under that texture name, with its dimensions, channel count, occupancy grid and brick pool. - `M.names() -> {string}` — every name `M.get` answers to, sorted. - `M.register(volume) -> Volume` — publish a wrapped volume under its own texture name. - `M.unregister(name) -> Volume?` — stop answering for a name and hand back the volume that held it, leaving its GPU resources alone. - `M.releaseTarget(name) -> boolean` — destroy a render target this module built for `render` and forget its size. - `M.sphere(opts) -> Volume` — procedural Volume filled with a soft sphere SDF. - `M.box(opts) -> Volume` — procedural Volume filled with a soft box SDF. - `M.noise(opts) -> Volume` — procedural Volume filled with FBM noise. - `M.raymarchShaderIdentity() -> string` — absolute identity of the shared raymarch shader. - `M.compositeShaderSource() -> string` — WGSL source for the post-process composite blending `scene_volume_rt` over the scene. - `M.FORMATS` — table of canonical format strings (`R8`, `R16F`, `R32F`, `RGBA8`, `RGBA16F`, `RGBA32F`). - `volume:destroy()` — release the GPU texture + occupancy. - `volume:writeBytes(bytes) / :writeFloats(floats)` — upload payload to the 3D texture (matches the texture's format). - `volume:fill(fn)` — CPU procedural fill via `fn(x,y,z) -> color`. - `volume:readback() -> bytes` — pull the texture's current contents back to Luau. - `volume:buildOccupancy({brickSize, sampler?})` — build an R8 occupancy companion for empty-space skipping. - `volume:buildSparse({brickSize?, threshold?, headroom?, capacity?, mode?})` — page the volume into bricks, keeping only the ones holding something. Yields while the brick count reads back — twice unless `capacity` says how many slots to take. - `volume:refreshSparse()` — re-page from the volume's current contents; two dispatches and no wait. - `volume:releaseDense()` — free the dense texture and render from the brick pool alone. - `volume:releaseSparse()` — free the brick pool and page table and go back to rendering the dense texture. - `volume:readSparseCounts()` — `{ bricks, wanted, overflow, capacity }` off the GPU; yields. - `volume:sparseStats()` — byte accounting: `denseBytes`, `poolBytes`, `pageBytes`, `residentBytes`, `ratio`. - `volume:fillCompute(shaderName, opts)` — drive a custom fill shader against this volume. - `volume:render(opts)` — dispatch the shared compute raymarcher into a 2D storage target, using `opts.entity` for AABB. A paged volume renders through its pool with no change at the call site. ## Usage ```luau local Volume = require("@builtin::systems.volumetrics.volume") -- Build a procedural noise cloud + occupancy. local cloud = Volume.noise({ name = "cloud", size = 64, scale = 4, octaves = 4, threshold = 0.55, }) cloud:buildOccupancy({ brickSize = 8 }) -- Render through the shared raymarcher; output to scene_volume_rt. cloud:render({ entity = entity(id), target = "scene_volume_rt", targetWidth = 1280, targetHeight = 720, density = 8.0, absorption = 1.2, emission = 0.9, color = { 1.0, 0.9, 0.7 }, maxSteps = 256, mode = "rgba", }) ``` The `Volume.component` is the per-entity wrapper around this module — it resolves the name it carries through `Volume.get` and dispatches `volume:render` from its `update` hook. ## Addressing a volume by name Every volume this module builds answers to its texture name, so anything holding only the name — an entity's `Volume` component, a second script, a tool — reaches the volume itself: ```luau local cloud = Volume.noise({ name = "cloud", size = 128 }) cloud:buildSparse({ brickSize = 8 }) cloud:releaseDense() -- somewhere else entirely Volume.get("cloud"):render({ entity = entity.find("cloud_host") }) ``` What the name alone cannot express is what makes this matter: the voxel dimensions, which channel holds density, and whether the volume is paged into a brick pool. A holder that reconstructs a wrapper from the name gets a dense volume that the paging may already have released. One volume owns a name at a time: the 3D texture, the page table and the brick pool all address it, so `create` over a name a live volume holds releases that volume first and says so in the log. `destroy` clears the entry. A wrapper around a texture another system owns is published with `Volume.register` and dropped with `Volume.unregister`, which hands the volume back so it can still be destroyed. The names live where this module is loaded — one set per Luau VM, alongside the raymarch target registry. `render` builds its `target` on the first call that names one and keeps its size, so repeated renders reuse it; `Volume.releaseTarget(name)` destroys that target and forgets the size, and the next render builds it again. ## Paging a volume into bricks A dense volume costs its bounding box. `buildSparse` splits it into bricks, keeps only the ones holding anything over `threshold`, and stores those in a brick pool; a page table one texel per brick says which slot each resident brick took, and 0 says the brick holds nothing. That page table is also the grid the raymarcher already skips empty space with, so paging and empty-space skipping are the same structure. ```luau local cloud = Volume.noise({ name = "cloud", size = 128, threshold = 0.6 }) local stats = cloud:buildSparse({ brickSize = 8, threshold = 0.01 }) print(stats.bricks, "of", stats.grid[1] * stats.grid[2] * stats.grid[3], "bricks") print(stats.poolBytes, "vs dense", stats.denseBytes) cloud:releaseDense() -- the pool is now the only copy cloud:render({ entity = entity.find("cloud_host") }) ``` Both passes run on the GPU with no readback between them, so a volume a compute shader rewrites keeps a matching page table for the cost of two dispatches: ```luau cloud:fillCompute(myFillShader, { extraResources = { { kind = "buffer", name = fillU } } }) cloud:refreshSparse() ``` The pool is fixed at the size the build measured, so a volume that grows past it loses the surplus bricks. `refreshSparse` asks the GPU for the counts without waiting on them and folds them in on a later call, so `sparseStats()` reports the overflow — and the log carries a warning naming the capacity to page at — a frame or two after the refresh that caused it. `buildSparse`'s `headroom` buys the slots up front for a volume that is expected to grow. What it costs: - A block in the pool is the brick plus a one-voxel apron on every side, so a resident brick costs `(brick + 2)³` voxels. At `brickSize = 8` that is 1.95x per resident brick, at 16 it is 1.42x. The apron is what makes a trilinear sample inside a brick read exactly what the dense volume would have given it. - A pool is always `rgba16f`: it has to be writable as a storage texture and sampled with filtering, and that is the format that is both. A scalar volume therefore spends more per resident voxel than it did dense (its density rides in the pool's red channel), so `sparseStats()` reports `denseBytes` and `residentBytes` side by side rather than assuming paging is the smaller of the two. - `refreshSparse` re-pages from the dense volume, so a volume that will keep changing keeps its dense copy; `releaseDense` is for one that is finished. - Classification reads one channel: alpha for a four-channel volume, red for a scalar one. A four-channel volume that carries density in red — the one `render`'s `mode = "density"` marches — has to say so, with the same word: `buildSparse({ mode = "density" })`. A `render` that names no `mode` marches a paged volume from the channel its pool was filled from, so the two agree without the caller repeating the choice. ## Notes - Volumes are GPU-owned; the Luau wrapper just holds the named handle. `volume:destroy()` releases both the texture and any occupancy companion. - The composite target `scene_volume_rt` is shared across every Volume entity in the scene, so multiple volumes composite into one image. Override `opts.target` if you need separate render targets per volume. - Occupancy grids use a brick size (default 8) — smaller = more precise skipping at the cost of grid memory. - The `vdb_to_zvol.importer` in this package produces dense voxel buffers consumable by `compute.writeFloatsTexture3D`, which can be wrapped via `Volume.wrap` (see the `zvol_viewer.scene/` demo for the full path).

# volumeDemos Shared orchestration backend for the `volumeDemos.package` showcase scenes. Owns the per-demo resource pack (3D texture, 2D render target, two uniform buffers) and the auto-registered composite, exposing a small public surface the demo scenes drive once per frame. ## Exports - `M.spawn({name, size, fillShader, fillUniformBytes}) -> ()` — register a new demo pipeline. Creates the 3D volume, the 2D storage target, registers the per-demo fill shader, ensures the shared raymarcher is registered, and auto-installs the depth-aware composite. - `M.destroy(name) -> ()` — release everything `spawn` created (texture + render target + buffers + composite registration). - `M.dispatchFill(name, fillUniforms) -> ()` — pack `fillUniforms` (Lua array of f32) into the demo's fill UBO and dispatch the fill shader against the volume. - `M.dispatchFrame(name, {camera, aabbMin, aabbMax, width, height, stepSize, maxSteps, alphaCutoff, densityGain, emissionGain}) -> ()` — pack camera + raymarch params into the shared 40-float UBO and dispatch the raymarcher into the demo's render target. - `M.NAMES(name) -> {volume, rt, composite}` — engine-side resource names for the demo (used by the `.zvol` viewer to bypass the helpers and upload voxels via `compute.writeFloatsTexture3D` directly). ## Usage ```luau -- From a demo scene at <pkg>/demo/<name>.scene/entrypoint.luau, -- `...` climbs to the package root, then resolves volumeDemos.module: local V = require("@builtin::systems.volumetrics.volumeDemos") -- onLoad: spin up the pipeline. V.spawn({ name = "smoke_plume", size = { 96, 192, 96 }, fillShader = SMOKE_FBM_WGSL, fillUniformBytes = 64, }) -- update(dt): repaint the volume + raymarch into the RT. V.dispatchFill("smoke_plume", { elapsed, density_amp, spread, ... }) V.dispatchFrame("smoke_plume", { camera = cam.get(), aabbMin = { -40, 0, -40 }, aabbMax = { 40, 160, 40 }, width = 1280, height = 720, stepSize = 0.9, maxSteps = 240, densityGain = 2.5, emissionGain = 1.0, alphaCutoff = 0.995, }) -- onUnload: free GPU resources. V.destroy("smoke_plume") ``` ## Notes - `dispatchFrame` packs 40 f32s into the raymarch UBO: `[0..16)` inv_view_proj (mat4 row-major), `[16..19)` aabb_min, `[19..22)` aabb_max, `[22..23)` step_size, `[23..24)` max_steps, `[24..25)` alpha_cutoff, `[25..26)` density_gain, `[26..30)` target_width/height + inv reciprocals, `[30..31)` emission_gain, `[31..40)` padding. - Pipeline state lives in module-locals keyed by `name` — the same module instance can host many demos in parallel, but two scenes must not pick the same `name`. - Each demo scene supplies its own fill shader; the raymarcher is shared (one shader name, one compiled pipeline). Replacing `fillShader` between frames is supported; replacing the raymarcher isn't (use the volumetrics package for that use case).

# volumeSequence Playback for a run of `.zvol` frames, through a ring of resident volumes. A sequence names its frames as VFS paths in playback order. It allocates `resident` 3D textures — one frame each — and nothing else: the frames the playhead is about to reach are read off the VFS on their own tasks and uploaded into whichever ring slot the playhead has travelled furthest past. So a 240-frame run and a 4-frame run cost the same on the GPU, and what a sequence costs is decided when it is opened rather than by how long it is. ```luau local Seq = require("@builtin::systems.volumetrics.volumeSequence") local frames = {} for i = 0, 239 do frames[#frames + 1] = string.format("/source/sim/plume_%04d.zvol", i) end local seq = Seq.open({ name = "plume", frames = frames, resident = 3, fps = 24 }) local s = seq:stats() print(s.residentBytes, s.allBytes) --> 3145728 251658240 ``` Drive it from a component, which is also what makes it visible: ```luau entity(id).component.add("VolumeSequence", { sequenceName = "plume", density = 6.0 }) ``` or by hand, if the caller already has a tick of its own: ```luau seq:tick(dt) local frame = seq:current() if frame then frame:render({ entity = entity(id), density = 6.0 }) end ``` The playhead advances once per engine frame however many callers tick the sequence — a component and a script, or two entities naming one run, share one playhead rather than each moving it. Every call still promotes the frame that has landed and starts the reads ahead of it. ## Residency Every frame of a run has the dimensions frame 1 declares, so one ring serves the whole run. `stats()` reports both what the ring holds (`residentBytes`) and what holding every frame at once would cost (`allBytes`); `residency()` says which frame each slot holds right now. The ring needs at least two slots for a multi-frame run — one holds the picture on screen while the other takes the read — and `open` raises the count to two if it was asked for less. `prefetch` says how many frames ahead of the playhead reads are started, and defaults to every slot the ring has spare — a slot left empty is a frame that gets waited on later. It is bounded by those same slots, so asking for more than fits reads as far ahead as it can. ## Keeping up A tick that wants a frame no slot holds yet keeps showing the frame it had and counts a `stall`. Playback is therefore never blank and never blocks, and `stats().stalls` is the number that says whether the ring and the prefetch distance are big enough for the rate: a run whose reads keep up shows a stall count that stops rising once the ring has filled. `errors` counts frames that failed to read or whose header disagreed with the ring. Each one is counted and logged once, with the path and what was wrong with it, and the run steps over it from then on rather than re-reading it every tick — so `errors` answers how many frames of the run are broken, and a mistyped path costs one line in the log rather than one per frame. The rate and the playhead are the sequence's, not a caller's: `setFps` keeps the frame the playhead stands on and the fraction of the way through it, and `seek` lands on exactly the frame asked for.

# volumetricLighting The settings behind the `VolumetricLighting` component, and the way to drive the scene's participating medium from a script — fog that rolls in over a cutscene, a quality slider that drops the step count, a room that fills with smoke. ```lua local volumetricLighting = require("@builtin::systems.volumetrics.volumetricLighting") volumetricLighting.set({ density = 0.06, anisotropy = 0.7 }) volumetricLighting.get() -- a copy of the settings in force volumetricLighting.active() -- whether the march is running volumetricLighting.clear() -- empty the air, keep the settings ``` `set` takes any subset of the fields; whatever is omitted keeps its current value. A `density` of 0 releases the render feature rather than marching a medium that scatters nothing, and `clear` is the named way to say the same thing. `sunDirection` tracks the scene's sun on its own. Pin it through `set` when a scene wants the medium lit from somewhere else. The module packs the settings into the `vol_params` buffer the passes read, so a change lands on the next frame without recompiling anything. ## Cost The march is the expensive part, and its cost is `steps` × (`shadowSteps` + the number of live scene lights) per marched pixel. Those are the two knobs to reach for on a budget: `steps` controls how cleanly a beam's edge resolves, and `shadowSteps` controls whether the sun's contribution is shadowed at all. Setting `shadowSteps` to 0 removes the per-sample visibility work entirely, which is the single largest saving available — at the cost of the shafts, since the medium is then lit evenly wherever it reaches.

# volumetricSky Orchestration backend for the `VolumetricSky.component`. Owns the GPU resources the cloud layer needs (cloud volume, raymarch render target, ping-pong TAA history pair, post-process composite) and exposes a small public surface the component drives every frame. ## Exports - `M.spawn(opts?) -> id` — spawn the singleton sky entity, or configure the one already standing, and return its id. Every weather parameter (`coverage`, `sunDir`, `sunColor`, `sunIntensity`, `windSpeed`, …) is settable at the top level or inside a `preset` sub-table; every `VolumetricSky` component field (`shaderVariant`, `cloudSize`, `taaBlendAlpha`, …) by its own name; `name` names the entity. Parameters left out keep their current values. An option that names neither raises, listing both sets, and leaves no entity behind. A weather parameter takes the shape its default carries — a number, or an array of that many numbers — and a value of another shape raises naming the parameter. - `M.configure(opts) -> boolean` — apply those same weather parameters and component fields to the sky already standing. Returns `false` when none has been spawned. - `M.defaultPreset() -> table` — a fresh copy of the weather configuration a sky spawned with no options carries. The copy is the caller's to edit. - `M.destroy() -> ()` — release every resource `spawn` created. - `M.fillClouds(noiseParams) -> boolean`: bake the cloud volume with the active `cloud_noise_*` shader when what it holds is not what `noiseParams` describe, and return whether this call baked. A change to a shaping parameter, the colour, the seed, the size, the variant or the render device bakes at once, and a write to the noise shader's `shader.wgsl` or `bindings.yaml` bakes once the program it compiles is resident; a change to the wind alone bakes at most `noiseParams.noiseRegenRate` times per second (default `M.DEFAULT_NOISE_REGEN_RATE`, `0` = every call the wind moved) on `noiseParams.clock` (default `getTime()`). `noiseParams` is the flat table produced by `VolumetricSky.component`'s `snapshotNoiseParams()`. - `M.volumeBakes() -> number`: how many times the cloud volume has been baked since the module loaded. - `M.dispatchRaymarch(renderParams, viewport) -> ()` — dispatch the active `cloud_raymarch_*` shader, blend with history via the TAA pass, and write into the composite target. `renderParams` is the flat table from `snapshotRenderParams()`. - `M.find() -> id?` — return the id of the active VolumetricSky entity, or `nil` if none has been spawned by `M.spawn`. - `M.despawn() -> ()` — despawn the entity `M.spawn` minted (does NOT free GPU resources — call `M.destroy()` for that). - `M.setCoverage(c) / M.setDensity(g) / M.setSun(dir, color, intensity?) / M.setWind(v)` — convenience setters that route through the active component's `public` fields. - `M.reprojectionDepthNdc(camPos, fwd, aabbMin, aabbMax, near, far) -> number` — the NDC z the TAA pass carries history through, taken from where the cloud stands along the centre view ray. A point at range `d` travels `translation / d` across the screen, so this is what keeps a translating camera's history landing on the part of the picture the cloud moved to. ## Usage ```luau -- The component drives this module automatically. Direct M.* use is -- rare; the canonical entry point is attaching VolumetricSky to an -- entity (or loading a preset and passing it to component.add): local sky = entity.spawn("sky") sky.component.add("VolumetricSky", {}) -- Driving the layer by hand instead, a frame at a time. Each dispatch -- creates the resources it needs at the size its own arguments name — -- the cloud volume from `noiseParams.cloudSize / cloudHeight`, the -- render target and composite from the viewport: local Sky = require("@builtin::systems.volumetrics.volumetricSky") Sky.fillClouds(noiseParams) Sky.dispatchRaymarch(renderParams, { width = 1280, height = 720 }) ``` ## Notes - `shaderVariant` selects which compute shader pair (`cloud_noise_v3 + cloud_raymarch_v3` / `cloud_noise + cloud_raymarch_v4` / `cloud_noise_v5 + cloud_raymarch_v5`). The module resolves the variant strings to absolute identities — relative paths don't work from a module body (the chunkname retains an `.init` segment that throws off the rsplit-to-parent step in `resolve_relative_dot`). - The cloud volume + history pair are RGBA16F — 256×128×256 ≈ 32 MiB VRAM. Drop to 192/96 via `cloudSize` / `cloudHeight` if tight. - The TAA pass is a single compute dispatch blending current sample with the previous frame's RT through the history pair. Set `taaBlendAlpha = 1.0` to disable while still going through the same code path (useful for benchmarking). - All public setters write through to `VolumetricSky.component`'s `public` fields, which means they replicate as Sync if the component is sync-enabled. Tweaks survive scene reloads as long as the same entity is alive. - `SkyPreset` and its defaults are declared here, and the component declares its `preset` field from them. `M.spawn` / `M.configure` read that same key set to tell a weather parameter from a component field, so a parameter added to the preset reaches the schema and the option routing in one edit.

# volumetricSky The cloud layer's render passes. The layer is raymarched for the camera being drawn, copied into the target the composite reads, temporally accumulated where there is a previous frame of that view to accumulate against, and laid over the frame behind whatever the depth buffer holds. Every pass but the accumulation runs for each view the frame draws, an environment cube's six faces included — which is what puts the clouds into reflections. The accumulation blends against the previous frame of the same view, and a cube face is rendered once with no previous frame, so it declares `environmentCapture = false` and the copy before it is what those faces read. `VolumetricSky` creates this feature; it is not meant to be created by hand.

# Package (asset type) A package is an asset — a `<Name>.package/` folder that bundles components, modules, scenes, presets, tools, and any other typed sub-assets under one identity, installed into a world as one unit. Packages are how authoring sets get distributed across worlds. A **library** is the addressable root scope content resolves under (`@builtin`, `@<name>`) — a whole world mounted into this one under a namespace, via `zeromind.installLib` / `world.installLibrary`. A package is an asset that lives inside such a root, and one library holds as many packages as its author put there. A package is installed by guid the way every asset is (`zeromind.install`, `zeromind.installAsset`). ## When to use one - You're authoring a coherent feature suite (character controller stack, voxel engine, UI kit, animation library) and want to ship it as one install target. - You want a stable identity worlds can install and resolve sub-assets against (`@<lib>::pkg.sub_module`). - You want versioning + a per-package metadata layer that travels with the contents. If you're authoring inside a single world and don't need a separate install lifecycle, you don't need a package — drop your `.component` / `.module` / etc. directly in `/zero/source/`. Packages are for cross-world distribution. ## Where it lives - Source: `/zero/source/.../<Name>.package/` (or, for legacy packages, a folder containing `package.yaml` at its root). - Identity: `<Name>` (the `.package` suffix collapses in identity — mirrors `.module`). A file at `<Name>.package/<sub>/init.luau` resolves under the same identity it would at `<Name>/<sub>/init.luau`. - Folder shape (container-style — open child set): - `package.yaml` — metadata (name, version, description). **Required.** - `README.md` — instance documentation. **Required.** - `.metadata` — agent-editable tags + free-form fields. **Required.** - Sub-asset folders: any combination of `.component/`, `.module/`, `.scene/`, `.preset/`, `.tool/`, `.toolbox/`, `.material/`, `.shader/`, etc. The validator permits unlisted children (`allow_unlisted: true`) by design — packages are open containers. ## How to create one ```luau asset.create("package", "<Name>") -- Creates: /zero/source/<Name>.package/ -- package.yaml (metadata scaffold) -- README.md (instance README template) -- `folder` places it in a subfolder of /source instead of the root: asset.create("package", "<Name>", { folder = "systems" }) ``` Then scaffold sub-assets inside it: ```luau local pkg = asset.resolve("<Name>.package") asset.create("component", "Spinner", { into = pkg }) asset.create("module", "util", { into = pkg }) ``` ## How it operates 1. **Registration.** Writing `package.yaml` indexes the package with the package registry. 2. **Install.** A package is read before it is installed: `zm ls-tree @<guid>` in the shell lists what it contains, and `zm show @<guid>:<path>` and `zm grep <pattern> @<guid>` read it, so candidates are compared by what they do. The one that fits is installed by guid, the way every asset is: `zeromind.install` from the shared library, or `tools.use("zeromind", "installAsset", <guid>)` inside the engine. The install writes the package's files (and the deps its `.refs` record) into this world's source, so its sub-assets resolve here. `tools.use("zeromind", "preview", <guid>)` reports what an install would write before it writes anything. 3. **Sub-asset resolution.** Because `.package` collapses in identity, sub-asset paths work whether the package folder is suffixed (`<Name>.package/<sub>`) or not (`<Name>/<sub>`). 4. **Updates.** `tools.use("zeromind", "updates")` lists installed content with upstream changes available; re-running the install pulls them. 5. **Metadata.** `package.yaml::name` overrides the folder stem if present (rare); `version` + `description` are surfaced in the package index. 6. **Hot reload.** Edits to sub-assets hot-reload as normal; edits to `package.yaml` reload package metadata. ## Discovery - `asset.list("package")` — every registered package. - `asset.inspect("<name>")` — sub-asset catalog, source, this type README. - `ls /zero/source/<Name>.package` lists the files the folder holds. - `library.list()` — the library assets registered in this world, including the ones packages under an installed library contribute. ## Authoring conventions - Name packages by their feature (`voxels`, `cameraRigs`, `dialogSystem`), not by their owner — the host library is the `@<lib>::` scope prefix. - Cluster cohesive sub-assets together. A "character controller" package should ship the controller component, its preset configs, its supporting modules, and a demo scene — not just the component. - Document the package's surface in `README.md` — what install installs, what each sub-asset does, how they relate. Agents reading `cat /zero/source/<Name>.package/README.md` get the full picture in one shot. - Version conservatively — breaking a package means breaking every world that installed it. Use semantic version bumps. ## Common pitfalls - **Missing `package.yaml`.** Without it the folder isn't recognized as a package even if it has a `.package/` suffix. - **Dependency installs.** A package that depends on another package must declare it (in `package.yaml::dependencies` once supported) and the parent world must install dependencies in order. - **Identity collapse.** Sub-asset paths skip the `.package` suffix. Don't include `.package` in `require` strings. ## Referencing inside a package: the `~` rule `~` means **this package**, resolved structurally from the referring file's own location. It is the only package-specific syntax. It is rename-safe (no package name is ever written) and library-agnostic (it never names the host library, so it survives `@builtin` being renamed or the package being mounted elsewhere). `~.tail` maps to `<pkg>.package/tail` from anywhere inside the package. | Referencing from | Use | Example | |---|---|---| | An asset/module **inside this package** | `~.tail` | `require("~.modules.myThing")`, `shader: "~.shaders.myThingBlend"` | | A **different package** under the same root | root-relative identity | `require("@root::systems.otherPkg.modules.x")` | | A **different library's** public API (a deliberate external dependency) | `@otherlib::…` | `require("@physics::modules.rigidbody")` | **Never write `@builtin::` (or any host-library name) to reference your own package** — that couples the package to the library it currently lives in and breaks if the library is renamed or the files move. Use `~.`. It works from Luau `require()` and from asset-reference strings in data files (a material's `mat.yaml` `shader:` field, a component's default asset ref); the engine resolves it against the referring file's own package. `asset.resolve` takes a `~.` name the same way, against the package of the file that calls it: `asset.resolve("~.materials.myThingStandard", "material")`. `~.` is the package-relative self-reference; `@root::` is the root-relative one (the library root, or the world source tree for non-library content). The `core/requiring-modules` guide covers the root model in full. A package that ships substantial functionality should include a `.guide` asset (a `<name>.guide/guide.md` folder anywhere under the package); the `guides` tool discovers it live, so collaborators find the system the way they find built-in guides. A package's test suite reaches the shared `Test` framework through its enclosing asset: `local Test = asset.containing(__FILE__).modules.shared` (`asset.containing` returns the deepest enclosing typed asset, the `.testSuite`, not the outer `.package`). ## Related types - Any of `.component`, `.module`, `.scene`, `.preset`, `.tool`, `.toolbox`, `.material`, `.shader`, `.style`, `.service`, `.importer` — packages bundle these. ## In the Inspector Package: the name, version and description package.yaml declares, how many things the package holds and what they are, and **Open** for package.yaml. A package.yaml that does not decode is listed under Problems with the decoder's error.

# Component (asset type) A component is a reusable behavior + data unit attached to an entity. The component file declares public fields, methods, and lifecycle callbacks; the engine instantiates one per attachment and drives its hooks every frame. Components are the primary unit of gameplay code — movement, AI, animations, UI panels, network sync, and per-entity state all live here. ## When to use one - You need per-entity state that survives across frames. - You need per-entity behavior driven by engine lifecycle hooks (`start`, `update`, `destroy`). - You want to expose configurable fields to designers / `entity.spawn` callers / scenes without code changes. If you need a global service that outlives any entity, use a `.service` instead. If you need a stateless pure-function library, use a `.module`. ## Where it lives - Source: `/zero/source/.../<Name>.component/` - Identity: `<Name>` (the `.component` suffix strips from the identity, but the folder retains the suffix on disk). - Folder shape: - `init.luau` (or `init.lua`) — declares the component. **Required.** - `README.md` — instance-level documentation. **Required.** - `.metadata` — agent-editable tags + free-form fields. **Required.** The structural contract is enforced by `asset.validate` against this type's `type.yaml`. ## How to create one Always start with the scaffold — it guarantees the folder shape matches `type.yaml` from the moment it's on disk: ```luau asset.create("component", "<Name>") -- Creates: /zero/source/<Name>.component/ -- init.luau (canonical declare + public block + hooks stubbed) -- README.md (instance README template) ``` After scaffolding, edit `init.luau` to fill in the actual fields and behaviour. The component registers as soon as the file is written — hot-reload is automatic; no engine restart required. ## How it operates 1. **Registration.** Writing `init.luau` into a `.component/` folder indexes the asset, mints a guid (in the `.meta` sidecar), and surfaces the type to the engine's component registry. Same-frame visible to `entity.component.add`. 2. **Attachment.** `entity(id).component.add(name, data)` (or the batch form `tools.use("entityOps", "addComponents", targets, { [name] = data })`) attaches an instance. The `data` table's fields override the component's declared defaults; omitted fields take the declared default. The `data` values are copied into `public.*` BEFORE `awake()` runs. 3. **Lifecycle.** The hooks run in this order: - `awake()` — same frame as attach, immediately after the data copy. Initialise state, set up native bridges, register with other systems. Other components on the same entity may not have awoken yet. - `start()` — the next frame. Every component on the entity has completed `awake()` by then, so cross-component setup is safe here. - `update(dt)` — same frame as `start()`, then every frame after, for as long as the gameplay clock is running. `fixedUpdate(dt)` runs beside it on the fixed timestep; `editorUpdate(dt)` runs while the clock is held. Which of them ticks is decided by `engine.mode` AND `engine.paused` together — see **Per-frame hooks and the two axes** below, and `man modes`. - `onDestroy()` — when the component is removed via `component.remove` or when the owning entity is despawned. Release resources, drop native ECS components. For the full list (`onEnable`, `onDisable`, `fixedUpdate`, `editorUpdate`, `onPropertyChanged`, `onAssetReload`, `onModuleReload`, `onBeforeSave`, …) and the timing details, **read `man components` — that is the canonical lifecycle reference. Generic component patterns (declare blocks, public fields, sync, ECS bindings, typed methods) all live there.** 4. **Public surface.** Fields declared in `public` are inspectable, editable in the inspector, and persist with the entity. Methods are declared on that same table with a colon — `typed function public:method(args)` — and every one of them takes the component ref as `self`, so a caller invokes it with a colon too: `entity(id).component.get(name):method(args)`. That is the one calling convention a component ref answers to: the reflection methods `ref:fields()` / `ref:getField(name)` / `ref:setField(name, value)` take the same spelling. A function the chunk declares at file scope (`function helper()`) is the component's own, reachable from its hooks under its bare name; `public` is what the ref carries. 5. **Settings and the Inspector card.** `type.yaml` names the component's settings schema as `settings: { schema: { instance: "init.luau", export: "public" } }`: the `public` table IS the schema, and its values live on each entity the component is on, so `asset.schema(componentRef)` answers the Field descriptors and no values file exists. On an entity the Inspector shows the component as a card whose generic `settings` section edits those fields as typed controls, each edit checked against the schema and written to the live component as one undoable operation. `type.yaml` also names `inspector: { instance: "inspector.luau" }`: a component that ships an `inspector.luau` beside its `init.luau` shapes its card — the file returns `{ sections, replace, order }` (see `assetTypes/README.md`, "`inspector:`"), each section's `build(ctx)` reading `ctx.component` (the live component) and `ctx.selection.id` (the entity). `SceneModule.component/inspector.luau` replaces `settings` with the bound module's declared inputs. 6. **Events.** An `events` block declares typed moments the component fires from inside — `events = { onHit = Event({ dmg = Field.number(0, NoSync) }) }`. The owner fires with `events.onHit:fire({ dmg = 10 })`; outside code subscribes through the proxy's subscribe-only facade, `entity(id).component.get(name).events.onHit:connect(fn)` (`:once` / `:wait` too). Pass `Sync` as the event's second argument to replicate a fire to other peers. Events let one component react to another's moments without polling. Full reference: **`man components`**, Events section. 7. **Hot reload.** Editing `init.luau` reloads the component definition; live instances are reconciled with the new declaration in-place, and each crosses one lifecycle boundary: `onDestroy()` ends the running life, then `awake()` and `start()` construct the one that replaces it, and `update()` resumes on the next frame. `public` and `private` values carry across; module-level `local`s belong to the chunk, so the new life starts with them at their initial values and its `awake`/`start` are what fill them in. Whatever the old life handed elsewhere — a subscription, a shared registration, a spawned entity — is released in `onDestroy`, the call that pairs with the construction about to run. An edit that leaves the source byte-identical crosses no boundary at all. A field carries its value across when something stated it: a key of the `data` table the instance was attached with (a scene record's data is that table), or a write made since. Every other field takes the edited source's declared default, so changing a default reaches every instance that left the field to it. An instance that states its own value for a field whose default the edit changed keeps that value, and the reload logs a `[HotReload]` warning naming the field, the entity, the value kept and both defaults. ## Discovery - `asset.list("component")` — every registered component (engine + world + libraries). - `asset.inspect("<name>")` — identity, scope, public fields, lifecycle methods, source path, and this type README. `detail` carries the component's declared `fields`, `methods`, and `events`, parsed from its source — `tools.use("assets","describe", "<name>")` renders it as markdown, so a component's events are visible without opening its source. - `ls /zero/source/<Name>.component` lists the files the folder holds. - `entity(id).component.list()` — components attached to an entity. ## Per-frame hooks and the two axes `engine.mode` (`"edit"` / `"play"`) and `engine.paused` (the gameplay clock) are independent, and the per-frame hooks read them separately: - `update(dt)` / `fixedUpdate(dt)` tick while **the gameplay clock is running** — whichever mode the engine is in. - `editorUpdate(dt)` ticks while **the clock is held** — because the mode is `edit`, or because play is paused — and also alongside `update` whenever edit is running. Which gives four states, and this table is the contract: | `engine.mode` | `engine.paused` | hooks that tick | |---|---|---| | `edit` | `true` | `editorUpdate` | | `edit` | `false` | `update`, `fixedUpdate`, `editorUpdate` | | `play` | `false` | `update`, `fixedUpdate` | | `play` | `true` | `editorUpdate` | Entering edit sets `paused = true` and entering play sets it back to `false`, so a session that never writes `engine.paused` only ever sees rows 1 and 3 — which is where the shorthand "`update` is the play tick, `editorUpdate` is the edit tick" comes from. It is the two default rows of a four-row table, not the rule. **Row 2 is the one that surprises.** With `engine.paused = false` in edit — set by hand, or by a tool that advances the gameplay clock to sample it, such as a `capture.collage` given a `duration` — `update` and `editorUpdate` both fire, once each, every frame. A component that declares both and routes them to one body therefore runs that body **twice per frame**: ```luau local function drive(dt) ... end function update(dt) drive(dt) end function editorUpdate(dt) drive(dt) end -- twice per frame in row 2 ``` Nothing reports the doubling; it reads as an animation running at twice its intended speed. To get exactly one tick per frame in every row, ask which axis you meant and answer it in the body: ```luau function update(dt) drive(dt) end function editorUpdate(dt) if engine.paused then drive(dt) end end ``` A scene `entrypoint.luau` declaring the same two hook names is dispatched by the scene loader, not by this table: there `update` needs play AND running, `editorUpdate` needs edit, and play-paused ticks neither. Read `man scenes` before carrying a component's rule to an entrypoint. ## Authoring conventions - Annotate public fields with `--!desc` so the inspector + LSP show documentation alongside the field name. - Declare lifecycle hooks (`start`, `update`, `destroy`) explicitly rather than relying on engine defaults — readers can see at a glance what runs when. - Keep `update(dt)` cheap — it runs every frame per instance. Move heavy work to `start`, to a service, or behind a job (`man jobs`). - For multi-entity batching, prefer `tools.use("entityOps", "addComponents", targets, components)` over a loop of `entity(id).component.add` — it takes many targets and many components in one call. ## Common pitfalls - **Naming.** The component identity is the folder stem with `.component` stripped (e.g. `<Spinner>.component/` → identity `Spinner`). Don't repeat the suffix in the identity passed to `component.add`. - **`init.lua` vs `init.luau`.** Either is accepted (it's a `one_of_group: "entrypoint"`), but `.luau` is canonical. Mixing both in one folder is invalid. - **`update` in edit mode.** It runs whenever the gameplay clock is running, and entering edit stops that clock — so `update` is idle in a freshly-entered edit session and ticks again the moment anything sets `engine.paused = false`. Use `tools.use("wld", "play")` to test update logic under the mode it ships in; `tools.use("wld", "edit")` to return. See **Per-frame hooks and the two axes** above. - **Renaming.** Renaming the folder changes the identity. Update every `component.add(<oldname>, ...)` call and the scene JSON entries that reference it. ## Related types - `.module` — pure-function library, no per-entity state. - `.service` — globally-scoped behavior, no per-entity state. - `.preset` — a captured component configuration, replayable into `component.add` without re-typing fields. - `.bundle` — entity hierarchies that may carry component attachments as part of the bundle. ## In the Inspector Fields: its execution order, hooks, and how many methods and events it declares, then the fields its `public` table declares as read-only typed controls holding their defaults, with **Check** and **Open** for its entry file. Instances: a link to every entity carrying it. What its card shows on an entity comes from the component's own `inspector.luau`.

# Module (asset type) A module is a reusable, stateless (or globally-stateful) Luau library that other code pulls in with `require()`. Modules are the unit of factored-out logic in a world: helpers, math libraries, format parsers, registries, shared algorithms. ## When to use one - You have logic you want to share across multiple components, services, scenes, or tools. - You have a registry, lookup table, or shared cache that callers all read from / write to. - You want a stable API surface (`M.fn`) other code can depend on. If the logic is bound to one entity, use a `.component`. If you want a named UI of fields rather than functions, use a `.preset`. ## Where it lives - Source: `/zero/source/.../<name>.module/` - Identity: the source path with `/` written as `.` and each segment's type suffix dropped — `/zero/source/demo/beats.module` is `demo.beats`, and `/zero/source/kit.package/beats.module` is `kit.beats`, the container's name kept and its `.package` suffix gone. `asset.create` returns this string, `asset.resolve(...).identity` reads it back, and the `require` forms below write it as `<identity>`. - Folder shape: - `init.luau` (or `init.lua`) — module body. **Required.** Returns a table (conventionally named `M`). - `README.md` — instance-level documentation. **Required.** - `.metadata` — agent-editable tags + free-form fields. **Required.** ## How to create one ```luau asset.create("module", "<name>") -- Creates: /zero/source/<name>.module/ -- init.luau (canonical `local M = {} … return M`) -- README.md (instance README template) -- `folder` places it in a subfolder of /source, and that subfolder -- leads the identity: `{ folder = "lib" }` produces `lib.<name>`. -- `into` authors it inside a resolved container (a package, a toolbox), -- whose own identity leads the same way: asset.create("module", "<name>", { folder = "lib" }) asset.create("module", "<name>", { into = asset.resolve("myPkg.package") }) ``` ## How it operates 1. **Registration.** Writing `init.luau` into a `.module/` folder indexes the asset and registers the module path with the resolver. 2. **Loading.** The first `require("<identity>")` runs the file's body and caches the returned table. Subsequent requires from anywhere return the **same table** — this is the shared-state lever for "module-local" registries. The body runs in an environment of the module's own: a name it declares without `local` is readable by the module's own functions and by no other script, so an export reaches its callers through the returned table and nothing else. Requiring a component's folder runs its chunk the same way, once, and its `update`, `start` and `public` stay in that environment; a component runs on an entity through `entity(id).component.add(name)`. 3. **Hot reload.** Editing the file re-runs the module body and folds what it returns into the table already loaded, so that table keeps its identity: a caller that captured it with `require` reads the new source's functions and values through the reference it already holds, and a later `require` returns that same table. A key the new source drops goes away; a key something set at runtime stays. The body's own `local`s start again from their initial values, so state that must outlive an edit lives in `modules.state()` (`guides { path: "core/scripting-and-tasks" }`). 4. **Identity resolution.** A module is reached by its `<identity>`, and where the caller lives decides which spelling of it resolves. For a module in a **world**, from a caller in that same world — another module, a component, a scene entrypoint, or an `execute` chunk: - `require("<identity>")` — the identity on its own, resolved against the caller's own root. It carries every folder segment, so a module created with `{ folder = "demo" }` is `require("demo.beats")`. - `require("@root::<identity>")` — the same root-relative resolution, written out. - `require(".sibling")` — a module in the same folder, by its bare name; `..name` steps up one folder, `...name` two. Logs, stack traces and hot-reload notices name that same module `@local.source.<identity>` — its key under the world source root. The `<identity>` tail of that key is the form above, so a trace reading `@local.source.demo.beats` is `require("demo.beats")`. For a module in a **library** — content under `/zero/source/libs/@<lib>/` — `@<lib>::` addresses that library's root from anywhere: `require("@builtin::modules.transform")`. Mounting a world as a library moves its files under such a root and flips the caller root to `@<lib>::`, which is what keeps the root-relative forms above resolving across the move. ## Discovery - `asset.list("module")` — every registered module. - `asset.inspect("<name>")` — public functions, source path, this type README. - `ls /zero/source/<name>.module` lists the files the folder holds. ## Authoring conventions - Return a single `M` table from `init.luau`. Top-level statements with side effects run on first require — useful for one-time initialization, dangerous if they touch the engine before it's ready. - Write exported functions as `typed function`, not `function`: the argument types are enforced at the call and the checker knows the signature, so a caller's mistake is reported where they made it. - Annotate exported functions with `--!desc` / `--!arg` / `--!return` / `--!example` so the LSP, `tools.list`, and `asset.inspect` surface them. A function with no `--!desc` is reachable only by someone already reading this file. - Declare a VALUE with `Field.<kind>(default, mode, description)`, and a table whose members are reached through a metatable with `--!members <TypeName>` — the latter gives the surface its accepted member set, its type, and its documentation from one table. See `guides { path: "core/authoring" }` for all four rules and `modules/api/engine/entity.module/members.module` for a worked member table. - Keep module state in module-local upvalues or in `modules.state()`. A name the body declares without `local` lives in the module's environment, which every run of the module shares: a function the previous run left in place reads what the reload assigned. Module-locals start again from their initial values with each hot reload; the table `modules.state()` returns is the same table on both sides of every reload. - Prefer focused modules over kitchen-sink modules. If two halves of a module have no shared state, split them. ## Common pitfalls - **Cyclic requires.** A `require` chain that loops will return the partial table (the half built before the cycle was detected). Design for one-way dependencies. - **Engine-time side effects.** Don't `entity.spawn` at module top-level — the engine may not be ready. Expose an `M.init` and call it from a scene entrypoint. - **Hot-reload + module-local cache.** A module that caches expensive computation in upvalues loses that cache on reload — fine for dev, but be aware in performance work. - **`init.lua` vs `init.luau`.** Either works, `.luau` is canonical. - **Writing one while play runs.** A `/source` write made while play is running lands on the play shadow, and a guarded play-exit discards it unless it is kept. One `.module` write shadows `init.luau`, the `README.md` and the `.metadata` together. `vfs.write(path, bytes, { durable = true })` skips the shadow, and `vfs.promotePlayShadow` keeps what is already on it — the `core/vfs` guide has the model and what each route costs. ## Related types - `.component` — for entity-bound state + lifecycle hooks. - `.service` — generates content you don't have yet (a mesh, sound, texture, …) via a metered provider. - `.tool` — for a single agent-callable function with a YAML schema. - `.package` — to group several modules + components + scenes into one shippable folder. ## In the Inspector Exports: what the module exports (name, kind, signature), parsed from its source without running it, with **Check** (the LSP's diagnostics for its entry file) and **Open**.

# Scene (asset type) A scene is a saved, loadable snapshot of an entity scenegraph inside a world: the entity hierarchy, transforms, components, lighting, and a declared `player` intent, plus an auto-discovered startup script. A scene is the unit of "level" or "screen" — a title screen, a lobby, a gameplay area, and an end-of-game summary each get their own scene. Scenes load as **layers**. One root scene is active at a time; additional scenes can load additively as overlays on top of it. Each scene owns its own multiplayer relay room. Switching scenes is a local call: the client that makes it leaves the old scene's room and joins the new scene's room. ## When to use one - You want a savable, re-loadable snapshot of an entity scenegraph plus its lighting and player setup. - You want a runnable starting state for a level or screen. - You want a scenes-as-data flow: `layers.load("lobby")`, edit live, then `layers.active:save()` to publish the edits back to the scene. For a reusable entity assembly (a prop, a vehicle, a UI panel) that many scenes spawn into themselves, use a `.bundle`. A scene is a whole-world snapshot; a bundle is one prefab that lives inside scenes. ## Folder shape A scene asset is a folder ending in `.scene`. Its identity is the folder name with the suffix stripped (`Lobby.scene` → `Lobby`). - `scene.json` — the scenegraph: `entities[]`, lighting, and the top-level `player` intent. **Required.** Written by the engine — a build, a save, a human dragging something in. Not hand-edited. - `build.luau` — what the scene is made of, written as code. Runs while you author; saving bakes its result into `scene.json` (see Building a scene). - `entrypoint.luau` — the startup script. Auto-discovered as a sibling of `scene.json`; define lifecycle callbacks here (see The entrypoint). - `<name>.<type>/` — an asset the build makes, authored here by `build.asset` (see The assets a build makes), alongside `.build.assets`, which records which `build.luau` produced each of them. - `prefabs/` — optional subfolder for scene-local bundles. - `scripts/` — optional subfolder for scene-local Luau helpers, required from `entrypoint.luau`. - `README.md` — optional per-scene notes about THIS scene. Author one only when there is scene-specific detail worth recording; the general scene model lives here, in `guides { path: "types/scene" }`. `scene.json` is version 7. It carries `"format": "scene"`, `"version": 7`, an `entities[]` array, and a top-level `"player"` string intent. ## The `player` intent Every scene declares how it handles players with a top-level `player` field in `scene.json`: - `"spawns"` — the scene carries an authored player setup, and the engine spawns a player for each connecting user from it. This is the default for a freshly created scene. - `"none"` — the scene spawns no player and authors no camera. Author your own Camera entity for the view (menus, UI-only screens, cinematics). ### How `"spawns"` works A `"spawns"` scene authors two things in `scene.json`: 1. A **PlayerPrototype** — an entity carrying the `PlayerPrototype` component, whose `body` field names a child entity (the body). The body carries an avatar (an `Asset` pointing at an avatar, e.g. `@builtin::avatars.humanoid`). A `CameraRig` child carries a `Camera` with a camera behavior (e.g. `orbital_follow`). The prototype root is authored `PrototypeOnly` so it exists as a clonable template. 2. A **PlayerSpawn** — an entity carrying the `PlayerSpawn` component, whose `prototype` field names the PlayerPrototype to instantiate and whose transform is where players appear. The scaffolded scene groups these under two organizing entities, `PlayerSetups` (holding the prototype) and `Spawns` (holding the spawn), so the authoring tree stays legible. On each user join — entering play, and each subsequent connect — the engine runs the spawn for that user: - It clones the PlayerPrototype subtree. - The clone root becomes the user's **internal identity** — a registry-only entity named `user` carrying the `UserIdentity` component. The identity is the anchor for who the player is; it has no world presence. - The body named by `PlayerPrototype.body` is adopted as the user's **avatar** — their visible presence in the world — and placed at the PlayerSpawn's world transform. Move the PlayerSpawn (or drag it in the editor) to set where players appear. - The `CameraRig`'s Camera follows the bound avatar. The player prototype is live in edit mode so you can see and adjust it. In play mode it is deactivated and marked internal — it stays a clonable source while each user gets their own clone. ### Reaching the live player The player surface lives on the root scene's registry, `layers.active.players`: - `players.localPlayer` — this client's player, a curated handle. - `players.localReady` — `true` once the local player's avatar is bound (or the scene opted its avatar out). - `players:list()` / `players:count()` — the players connected in this scene's room. - `players:get(userId)` — a connected player by account id. - `players:ownerOf(avatar)` — the player whose body is a given avatar. A player handle is a data record about who the player is, plus a link to their body: - `player.avatar` — the body's entity ref. Act on the body through it (`player.avatar.position`, raycasts, rig access). - `player.isLocal` — whether this client owns the player. - `player.ready` — whether a live avatar is bound. - `player.userId` / `player.identity` / `player.displayName` — the player's account identity. - `player.avatar = ref` — bind a live body entity as the player's body (assign nil to clear). To use a `.bundle` / `.avatar` asset, spawn it into the world first, then bind the resulting entity. The player's identity entity is the internal anchor; the handle exposes the body (`player.avatar`) and the player's data, and reaching for `.entity` / `.id` / `.component` raises with a redirect to `player.avatar`. ## Building a scene `build.luau` declares what the scene holds. It is ordinary engine code — `entity.spawn`, `component.add`, transform writes, `for` loops, a `require` of a module that computes a layout — and what it creates is what ends up in the scene. ```lua function content() for i = 0, 4 do local crate = entity.spawn("crate") crate.localPosition = { i * 2, 0.5, 0 } crate.component.add("Model", { model = asset.resolve("cube", "mesh") }) end end ``` The build runs while you author, in **edit** mode, and what it created lands in the live scene. Saving the scene bakes it into `scene.json`, which is what makes the entities durable content: a human can select and inspect them, they replicate, and they are there when the world opens without `build.luau` running at all. Play mode never runs it — in play the entities load from `scene.json` like anything else the scene holds. **Editing the file rebuilds the scene in place.** Writing `build.luau` is the signal; there is nothing to reload, and the rebuild saves. An entity keeps the id it had, so a reference to it — a spawn naming a prototype, another entity's component naming this one — survives every rebuild. Add a crate to the loop and the other four do not move. The write returns as soon as the file is on disk and the rebuild finishes behind it, so what it landed comes back as a notice: ``` [info] a write to build.luau rebuilt the scene { content="6", scene="scenes.main", editorOnly="1" } ``` An edit saved while a rebuild is still running is an edit to code that rebuild has already read, so it gets a rebuild of its own once that one finishes, and says which it was: ``` [info] a write to build.luau arrived during a rebuild and was rebuilt after it { content="7", scene="scenes.main", editorOnly="1" } ``` Several edits saved during the same rebuild get the one run after it, which reads the file as it stands then. A scene that declares a build also runs it when it **loads** in edit, so the scene shows what the code says now, including an edit made while the world was closed. A load whose `.build.state` records the current script reads `scene.json` as the build and runs nothing. When the load does run it, a world's own scene saves the result, unless the scene holds unsaved edits: those stay the author's to keep or discard, and the file write waits for their save. A library scene (one under `/libs/`, the `@builtin` library included) is read by its load: the build lands in the live scene and its files stay as the library states them. Play reads `scene.json`, so it comes up without that build until `:build()` saves it. ### The two surfaces - `content()` — the scene's own entities. Present in play. - `editorOnly()` — entities present while authoring and absent in play. They carry the `EditorOnly` participation the runtime deactivates and hides when play begins, and they are baked too, so a collaborator opening the world sees the same ones. Anything an author needs to see and a player does not goes here — an alignment guide, a spacing marker, a debug volume. Both are optional. A scene with neither has no build. ### What a rebuild touches Only the entities the build itself placed. Anything a human dragged in, or a tool placed directly, is invisible to a rebuild and never moved or removed. Within what it owns, the build is authoritative: an entity the code stopped emitting is despawned, a component it stopped adding is removed. The build writes its own name and each entity's place in the hierarchy onto the entity, and `scene.json` records that pair beside the entity's name and transform — which is why a rebuild after a reload, in a later session, lands on exactly the entities it placed the first time without anything having been remembered. ### When the file did not change `:build()` runs the build against what it resolves right now — for the case where something the builder *reads* changed and the file did not: a module it requires, an asset it resolves. After a write it is unnecessary; the write has already run it. ```lua layers.active.asset:build() -- rebuild and save layers.active.asset:build({ save = false }) ``` It answers with how many entities each surface placed — `{ content = 6, editorOnly = 1 }` — and with nil for a scene that has no `build.luau`. Called while a build for that scene is already running, it returns at once with `{ inFlight = true, scene, trigger, startedBy, message }` naming that build instead of starting a second one; `inFlight` is what tells the two answers apart. An operation the bake cannot hold is **refused** rather than applied. A build accepts what its records carry: entity lifecycle, transforms, hierarchy, names, visibility, active state, participation mode, network scope, attributes, and components with their public data. Anything else raises while the build runs, naming the operation and the build it answered to, instead of leaving the scene depending on something no reload brings back. When that happens the previous build stays exactly as it was. A runtime resource a build needs — a mesh it generates, a texture it writes — is made as an asset instead, below. Everything else a refusal names goes outside the builder, or into a component the builder attaches, which runs at play. ### The assets a build makes `build.asset(type, name, produce)` makes an asset the build's content names — a mesh a lathe produces, a texture a pattern writes. `produce` takes nothing and returns the creation parameters `asset.create` takes for that type; the call returns the asset's `AssetRef`, which is what a component field holds. ```lua function content() local tower = build.asset("mesh", "tower", function() local positions, indices = lathe(profile, 16) return { positions = positions, indices = indices } end) entity.spawn("tower").component.add("Model", { model = tower }) end ``` The asset is authored inside the scene's own folder — `Lobby.scene/tower.mesh` — and the scene records which `build.luau` produced it in `.build.assets`. So `produce` runs when the build script changed and never otherwise: a reload re-runs the build and reuses the asset, while editing the profile re-authors `tower.mesh` **in place**, keeping its path and its guid. That is what holds the reference — `scene.json` names assets by guid. For a part of a scene that several scenes place — or that one scene places many times with different settings — see `guides { path: "types/sceneModule" }`, which is the same idea with declared `inputs` a placement sets. `guides { path: "core/scenes-as-code" }` covers both. ## The entrypoint `entrypoint.luau` is the scene's **runtime** — what happens while the game is being played. What the scene *is* comes from `build.luau` above; nothing here has to construct it. `entrypoint.luau` is auto-discovered as a sibling of `scene.json`. The loader runs it in a fresh environment and folds its top-level `function name(...)` declarations into the matching scene lifecycle event. Define the callbacks the scene needs — the recognised set: - `onLoad()` — the scene loaded in **play** mode. Gameplay-start hook, runs on every connected client. - `onHostLoad()` — the scene loaded in **play** mode, on the client that owns the scene's synced content (the relay room creator; offline and single-player count as host). Fires once. Entities spawned here — and anywhere downstream — become synced automatically, so this is where shared world content is spawned: it runs on exactly one peer, and every other peer receives the entities from the relay snapshot. - `onEditLoad()` — the scene loaded in **edit** mode. Custom per-scene authoring tools. Writing this file while the scene is loaded in edit mode reloads its layer, so the new `onEditLoad` runs after the previous load's `onUnload`. - `onUnload()` — before the scene's entities are torn down. Extra cleanup; engine-managed teardown (entity despawn, resource release) runs separately. - `update(dt)` — per frame in play mode, while the gameplay clock runs. - `editorUpdate(dt)` — per frame in edit mode, running or paused. Play-paused ticks neither hook. The scene loader partitions these strictly on mode; a component's hooks of the same name read `engine.paused` instead, so the two rules differ — `man components` has that table. - `localPlayerReady(player)` — once the local player's avatar is ready. Gameplay wiring for the local player: attach state, bind UI, swap the avatar. The avatar is already placed at the PlayerSpawn, so this is for behavior, not positioning. - `playerJoined(player)` — a remote player joined this scene's room. Use for per-player UI (nameplates, greetings). The local player arrives through `localPlayerReady`. - `playerLeft(eid)` — a remote player left this scene's room. Receives the departed entity's id (the entity is already despawned). Each scene's entrypoint runs in its own environment, so two scenes can declare the same top-level names without clashing. Keep each callback short; push per-scene logic into `scripts/` modules and `require` them. ## Layers Scenes load through the `layers` namespace. - `layers.load(ref, opts?)` — load a scene. `ref` is a scene AssetRef or identity string. Without `opts.additive`, this is a **root** load: it replaces the current root scene. `opts`: - `additive` — load as an overlay alongside the current root instead of replacing it. - `persistent` — keep an additive overlay mounted across mode flips and root swaps. - `origin` — a `{x, y, z}` world offset for the overlay's entities. - `deferred` — spread the entity spawn across frames (for thousand-entity scenes); the load signals complete only once the scene has settled. - `layers.active` — the current root scene proxy (nil before any load). - `layers.list()` — every loaded layer (root + additive overlays). - `layers.find(ref)` / `layers.is_loaded(ref)` — look up a loaded layer by ref. - `layers.onLoad(cb)` / `layers.onUnload(cb)` / `layers.onBeforeLoad(cb)` — subscribe to every scene load / unload (distinct from a scene proxy's own `:on_load` / `:on_unload`, which fire only for that scene). A root scene carries the `players`, `camera`, and `settings` surfaces; additive overlays inherit the root's players and lighting and do not carry their own. ### The scene proxy `layers.active` (and `layers.find` / `layers.list` entries) returns a scene proxy: - Fields: `.guid` (canonical identity), `.name`, `.path`, `.asset`, `.additive`, `.visible`, `.state` (`"loading"` → `"loaded"` → `"ready"` → `"unloading"`), `.ready`, `.players`, `.camera`, `.settings`, `.entrypoint`. - `.camera` — a proxy over the scene's primary camera entity; read and write the live Camera component through it, and `.camera.behaviors._list()` enumerates the available camera behaviors. - `:save(opts?)` — publish the live scenegraph to canonical `scene.json` (see Saving). A convenience that forwards to the scene ASSET's `save` (`self.asset:save(opts)`): the proxy is the runtime view and owns nothing on disk, so persistence lives on the asset. - `:reload()`: unload and reload the same scene (a runtime respawn). A scene with a `build.luau` runs its build again and saves the scene, so `scene.json` holds what the reload built and the next play loads it; a library scene's rebuild lands in the live scene alone. - `:clearDirty()` — discard unsaved edits and respawn from canonical. Forwards to the scene asset's `discard` (`self.asset:discard(opts)`). - `:sceneAsset()` — the durable `AssetRef<scene>` this live layer maps to. - `:promoteDirty()` / `:writeDirty()` / `:hasDirty()` — manage the dirty overlay directly. - `:unload()` — despawn the layer and free its slot. - `:onReady(cb)` — fire once the scene reaches `ready` (latched: a late subscriber fires immediately). ## Multiplayer & relay rooms Every scene has its own relay room, keyed by the world guid, the boot profile, the mode, and the scene guid. Switching scenes with `layers.load` runs a room transition — leave the old room, tear down, join the destination scene's room — all on the client that made the call. Teardown of the old scene never leaks to peers in the new one. In play mode connected to a relay, the scene's entities are host-authoritative: only the room creator instantiates them (marked synced), and every joiner receives them from the relay snapshot rather than spawning their own. This is why shared world content belongs in `onHostLoad` — spawning it in `onLoad` would run on every peer and duplicate it. Each player's own avatar spawns per-user from the PlayerSpawn on every client. Edit mode and offline boots spawn locally and never auto-sync. ## Loading and saving **Loading** — `layers.load(ref)` for a root scene: 1. Despawns the current root scene's non-persistent entities. 2. Instantiates the entities from `scene.json` (including the player prototype and spawn for a `"spawns"` scene). 3. Applies lighting and sky. 4. Runs `entrypoint.luau` and fires lifecycle callbacks as their events occur. **Saving** — persistence lives on the scene ASSET. `sceneRef:save()` writes the scene's current state back to canonical `scene.json`; when the scene is loaded it captures the live scenegraph, and when it is not loaded it promotes the pending dirty overlay. `layers.active:save()` is the convenience form — it forwards to the active layer's mapped asset (`layers.active.asset:save()`). Writing canonical is the only path that does so: edit ↔ play cycles never save, so hitting play to test never overwrites the scene. Between saves, live edits accumulate in a per-scene dirty overlay that the loader merges on top of canonical. `sceneRef:discard()` throws the overlay away and restores the saved state — deleting the overlay, and respawning the live layer when the scene is loaded. It works whether or not the scene is loaded, because the overlay belongs to the asset, not the runtime layer. `layers.active:clearDirty()` forwards here. Saves capture entity data, not the entrypoint script — edits to `entrypoint.luau` and `scripts/` are always preserved. `opts.to` on `save`/`discard` targets a different scene ("save as"), taking either a bare name (`"level_2"`) or a full VFS path. **What a record states.** A component's record in `scene.json` holds the fields that instance's configuration states: the keys of the table it was added with (a `build.luau` call's own init table, or the record it was loaded from) and every field written since, through code, a tool or the inspector. A field nothing set stays out of the record and takes the component's declared default each time the scene loads, so changing that default reaches every entity that left the field to it. To have an entity follow the default again, remove the field from its record. A field set on a live entity after its record was made is a change the scene review names (`sceneAuthoring.changes` lists it as `+ Type.field = value`), and accepting it writes the field into the record. A scene saved before records were stated this way holds every field; those values load and save as they are, since a stored value cannot be told apart from one an author chose, and a hot reload that changes a default one of them shadows names it. A stored field the component no longer declares is named by a `[component.add]` warning when the record loads and left out of the next save. ## Lighting Lights are ordinary entities: a directional light, an ambient light, and a sky are entities carrying `Light` / `ProceduralSky` components in `scene.json`. Author, move, and tune them like any other entity. If a scene is missing any of {directional light, ambient light, sky}, the loader adds a temporary default for the absent one so a scene never loads dark or skyless; the defaults are never saved into the scene. Scene-level lighting state that is not an entity — clear color and sky descriptor — lives under `layers.active.settings.lighting` (`clear_color`, `sky`); writing those fields updates the engine live. ## Modes Scenes load in both edit and play mode (see `man modes`). The mode governs behavior: - An entrypoint's `update(dt)` runs in play while the gameplay clock runs, and its `editorUpdate(dt)` runs in edit. A component's hooks of the same name follow the clock rather than the mode — `man components` has the four states that produces. - Player prototypes are live in edit and deactivated + internal in play, where each user instead gets a clone. `EditorOnly` entities follow the same edit-live / play-internal toggle. - `onLoad` / `onHostLoad` fire on a play-mode load; `onEditLoad` fires on an edit-mode load. ## How to create one ```luau asset.create("scene", "<Name>") -- Creates /zero/source/<Name>.scene/ with: -- build.luau where this scene's content is written. A new one starts -- with the sun / ambient / sky lighting and the "spawns" -- player setup, commented for editing — everything else the -- scene comes to hold is written here too. -- scene.json what that code produces, baked. -- entrypoint.luau commented lifecycle-callback stubs. ``` A freshly created scene has a player with a following camera on the first press of play. It carries no ground, so the first thing to author is what the player stands on — a `content()` line away in `build.luau`. ## Discovery - `asset.list("scene")` — every registered scene. - `asset.inspect("<name>")` — entity count, lighting summary, source path, and this type README. ## Authoring conventions - Keep static layout in `scene.json` — geometry, lights, props, the player prototype and spawn. - Keep dynamic setup in `entrypoint.luau` lifecycle callbacks — event listeners, gameplay state, services attached to spawned entities. - Spawn shared, synced world content from `onHostLoad`; spawn each player's own gameplay wiring from `localPlayerReady`. - Set the world's startup scene in `/.world_settings` — that is the scene the world boots into. - Y is up; +Z is forward. - Renaming the folder changes the scene's identity — update every `layers.load` callsite and the world's startup-scene field. ## Related types - `.bundle` — reusable entity hierarchies that scenes spawn into themselves. - `.service` — cross-scene background work that outlives a scene load. - `.world_settings` — picks the startup scene and world-level renderer / physics options. ## In the Inspector Selecting a scene opens its **Scene** section, read from `scene.json` without loading the scene: how many entities it holds, how many of them are roots (and their names), how players join and the file's format version when it states them, and whether the scene has edits not yet saved.

# Compute shader (asset type) A `.computeShader` is a GPU compute (GPGPU) program with a **zero-scaffolding** authoring contract: you write only `@compute fn main(...)` plus a declarative `bindings.yaml`, and the engine GENERATES every `@group/@binding` declaration — storage buffers, textures, samplers, storage textures, and a `params:` uniform — expands `#include`s, naga-validates the result, and registers it. The single source of truth for the binding layout is `bindings.yaml`; there are no hand-written `@binding` lines to keep in sync. > New to how content lives across the CPU and GPU? Read `core/resource-model` > first: it covers the asset / CPU / GPU forms every resource takes, the upload > path, and how compute fits the performance ladder (`core/performance`). ## Folder shape A `.computeShader` is a folder asset: ``` <name>.computeShader/ shader.wgsl # the WGSL body — only `@compute fn main` + helpers (required) bindings.yaml # the binding + params schema (required) README.md # a one-paragraph description of THIS shader (required) .metadata # asset metadata ``` Each file carries a committed `.meta` sidecar pinning its stable guid. The identity is `<name>` (the `.computeShader` suffix strips). Resolve the shader once, then dispatch through the reference: ```lua local carve = asset.resolve("carve", "computeShader") carve:dispatch({ buffers = { heights }, workgroups = { 64 } }) ``` `asset.resolve` is what records your script's dependency on the shader, so the shader travels with the world your script lives in — packed, pulled, or published. It compiles on first dispatch; there is no setup step. ## bindings.yaml One ordered `bindings:` list — the binding index is the list position (0, 1, 2, …) — plus an optional `params:` scalar-uniform block. `kind` defaults to `buffer` when omitted. ```yaml bindings: - { name: clouds, kind: storage3d, format: rgba16f } # texture_storage_3d<rgba16float, write> - { name: occupancy, kind: texture3d } # texture_3d<f32> - { name: smp, kind: sampler } # sampler - { name: out_tex, kind: storage2d, format: rgba16f } # texture_storage_2d<rgba16float, write> - { name: U, kind: buffer, access: read, element: f32 } # var<storage, read> U: array<f32> - { name: Uniforms, kind: buffer, access: read, element: VolumeUniforms, array: false } # var<storage, read> Uniforms: VolumeUniforms params: - { name: scale, type: float, default: 2.0 } - { name: count, type: int, default: 0 } ``` | `kind` | extra keys | generated WGSL declaration | |---|---|---| | `buffer` (default) | `access: read \| read_write` (default `read_write`), `element: <wgsl type>` (default `f32`), `array: true \| false` (default `true`) | `var<storage, ACCESS> NAME: array<ELEMENT>;` (or `… NAME: ELEMENT;` when `array: false`) | | `texture3d` | — | `var NAME: texture_3d<f32>;` | | `texture2d` | `sample: float \| uint \| sint` (default `float`) | `var NAME: texture_2d<f32>;`, or `<u32>` / `<i32>` for the other classes | | `texture_depth` | — | `var NAME: texture_depth_2d;` — the one sampled shape with no `sample:` of its own; a depth-aspect channel (`@scene.depth`) binds here, not through `texture2d` | | `storage3d` | `format`, `access: write \| read \| read_write` (default `write`) | `var NAME: texture_storage_3d<FMT, ACCESS>;` | | `storage2d` | `format`, `access: write \| read \| read_write` (default `write`) | `var NAME: texture_storage_2d<FMT, ACCESS>;` | | `sampler` | — | `var NAME: sampler;` | | `textureCubeArray` | — | `var NAME: texture_cube_array<f32>;` — the environment cube array, bound to `@scene.environment` by a render feature's pass | | `vertices` | `access: read \| read_write` (default `read_write`) | `var<storage, ACCESS> NAME: array<u32>;` plus the mesh's layout uniform and the `zeroVertex*` accessors | | `acceleration_structure` | — | `var NAME: acceleration_structure;` (hardware ray tracing) | - **Eight storage buffers per shader** — the WebGPU guarantee, and what the engine asks every device for. `buffer` and `vertices` bindings spend that budget; textures, samplers, storage textures and `params:` do not. An `acceleration_structure` brings engine-owned buffers with it — **three** where the device has hardware ray query (the tables `zeroHitSurface` / `zeroMaterial` read) and **five** on the software backend (those, plus the scene triangles and the hierarchy over them) — so a traced pass has three slots of its own on the backend web users get. Over budget, the engine refuses the shader and logs which buffers it declared. - A `buffer` defaults to `array<ELEMENT>`. Set `array: false` for a single NON-array storage buffer — a struct binding (e.g. `element: VolumeUniforms`, `array: false` → `var<storage, read> NAME: VolumeUniforms;`). The struct definition itself lives in `shader.wgsl`, not `bindings.yaml`. - **A sampled texture binds under the class its FORMAT puts it in**, so `sample` states that class and the generated `texture_2d<..>` matches it. An integer channel — `@scene.material`, which is `Rgba8Uint` — is `uint`; everything else is the `float` default. Bind an integer channel to a `float` slot and the engine refuses the pass and names both classes, because the device would otherwise reject the bind group and invalidate the pipeline. - A storage texture's `access` decides both the generated declaration and the layout the engine registers, so the two always agree. It defaults to `write`, the access every target grants every storage-compatible format. `read_write` is granted on `r32f` alone; declaring it on any other format is refused at the compile, with the reason on `compute.programState`. A pass that needs to read a wider target reads it through a separate `texture3d` / `texture2d` binding, or through a second storage binding declared `read`. - `format` tags (storage textures): `r8`, `r16f`, `r32f`, `rgba8`, `rgba16f`, `rgba32f`. Storage textures only allow storage-compatible formats (`r32f` / `rgba8` / `rgba16f` / `rgba32f`). - An `acceleration_structure` binding resolves to the scene acceleration structure when the shader runs as a render-pass with ray tracing enabled (`renderer.setRaytrace(true)`). Declaring one makes the engine emit the helper functions the body can call: `zeroTraceClosest(origin, dir, t_min, t_max) -> ZeroHit` (struct: `hit, t, instance, primitive, bary, attr`, where `instance` is the hit object's render slot), `zeroTraceAny(origin, dir, t_min, t_max) -> bool` (terminate-on-first-hit, for shadow/occlusion rays), and `zeroHitSurface(hit) -> ZeroSurface` (`normal, uv, material`) + `zeroMaterial(index) -> ZeroMaterial` (`base_color, emissive, metallic, roughness`) to shade what a hit landed on. The same surface works on both backends — hardware ray query and the software/compute path — so the shader doesn't branch on `renderer.raytraceCapability()`. Bind it from a render-pass, which is where the scene acceleration structure exists. `params:` declares scalar uniforms (`type: float | int`). When non-empty the engine appends a `ZeroComputeParams` uniform as the trailing binding (read it in WGSL as `params.<name>`) and tunes its values at runtime with `shaderRef:setParam(prop, value)`. ## The kernel ABI: `columns`, counted and appended outputs, `specialize`, `determinism` One shader asset can serve many shapes of data. `bindings.yaml` states the shape with these extensions, and the engine's compile path lowers them to the plain bindings and WGSL above (`modules/compute_abi`, whose README has the whole surface): ```yaml workgroup: 64 # the size the body declares (default 64) bindings: - { name: cols, kind: columns, access: read } # a table of columns in one binding - { name: src, kind: buffer, access: read, element: T } - { name: dst, kind: buffer, access: read_write, element: T, count: "len(src)" } - { name: hits, kind: buffer, access: read_write, element: vec4<f32>, capacity: "n * 4", append: true } specialize: - { name: T, options: [f32, i32, vec2<f32>, vec3<f32>, vec4<f32>] } - { name: FOLD, options: [add, min, max] } - { name: BODY, kind: source, default: "0.0" } # text that replaces the identifier determinism: exact # or float_order ``` - A `buffer` of a scalar or vector element gets `NAME_get(i)`, `NAME_len()` and (unless `access: read`) `NAME_set(i, v)`. A `vec3` element is stored as a flat scalar array, so a column of vec3 keeps its 12-byte stride. `zero_n()` is the dispatch size. - `columns` binds a table of columns (`compute_abi.table`) as one storage buffer; `cols_f32(c, i, lane)`, `cols_vec3(c, i)` and the rest read it. - `count` / `capacity` give the size, in elements, of an output the run allocates. `append: true` adds a live counter binding, `NAME_count`, with `NAME_store(slot, v)`, `NAME_set_count(n)` and, under `determinism: float_order`, `NAME_append(v)`. - `specialize` parameters are compile-time: type options become `alias NAME`, named options become `const NAME` and `NAME_<option>`, `kind: source` text replaces the identifier. Each specialization compiles once. - `determinism: exact` refuses `NAME_append`: an exact output is written at ranks the kernel computed. Dispatch such a shader through its program: ```lua local variant = shaderRef:program():variant({ specialize = { T = "vec3<f32>" } }) local run = variant:run({ n = 64, inputs = { src = srcBuffer } }) local values = require("@builtin::modules.compute_abi").read(run.outputs.dst.buffer) ``` `shaderRef:program()` answers the same program, with the variants it has lowered and compiled, for as long as `shader.wgsl` and `bindings.yaml` hold the bytes they held when it was made. A write to either file, from any origin (`vfs.contentVersion` moves), makes the next `program()` a new one, and a dispatch through it runs the new text. The schema `getBindings()` answers is the one the program was made over. ## Writing a mesh's vertices — `kind: vertices` A mesh's vertices are raw bytes on the GPU, and the engine stores a mesh in one of two layouts: 92 bytes a vertex with every attribute a `f32`, or 32 bytes a vertex with normals and tangents quantized to bytes and UVs to shorts. Which one a mesh got depends on the mesh. Declare `kind: vertices` and the shader is handed both the buffer and the layout that mesh is in, plus accessors that address it: ```yaml bindings: - { name: verts, kind: vertices } # the mesh, writable - { name: heights, kind: buffer, access: read, element: f32 } params: - { name: lift, type: float, default: 1.0 } ``` ```wgsl @compute @workgroup_size(64) fn main(@builtin(global_invocation_id) id: vec3<u32>) { if (id.x >= zeroVertexCount()) { return; } var p = zeroVertexPosition(id.x); p.y = p.y + heights[id.x] * params.lift; zeroSetVertexPosition(id.x, p); zeroSetVertexNormal(id.x, vec3<f32>(0.0, 1.0, 0.0)); } ``` ```lua carve:dispatchOnVertices({ model = meshRef:handle().guid, buffers = { heights }, workgroups = { 64 } }) ``` The accessors, over a binding declared `read_write`: | read | write | |---|---| | `zeroVertexCount() -> u32` | — | | `zeroVertexPosition(i) -> vec3<f32>` | `zeroSetVertexPosition(i, vec3<f32>)` | | `zeroVertexNormal(i) -> vec3<f32>` | `zeroSetVertexNormal(i, vec3<f32>)` | | `zeroVertexTangent(i) -> vec4<f32>` | `zeroSetVertexTangent(i, vec4<f32>)` | | `zeroVertexUv(i) -> vec2<f32>` | `zeroSetVertexUv(i, vec2<f32>)` | | `zeroVertexColor(i) -> vec4<f32>` | `zeroSetVertexColor(i, vec4<f32>)` | | `zeroVertexNodeIndex(i) -> u32` | — | | `zeroVertexWord(i, offset) -> u32` | — | `access: read` gives you the readers only — a `var<storage, read>` binding cannot be assigned to, so the writers are left out rather than handed to you broken. Once a pass has written a mesh's vertices, the engine bounds that mesh by what the pass wrote: it reduces the vertices to an AABB every frame and culls against that, so geometry a pass moves across the scene needs no oversized placeholder to survive culling. `renderer.mesh.boundsSource(guid)` reports `"compute"` for such a mesh and `"geometry"` for one still bounded by what it was created with. A shader declares at most one `vertices` binding; the accessors are generated for that mesh. It can be declared anywhere in the list — the mesh is bound by the dispatch, and `buffers` fills the other storage slots in declaration order. The tangent's `w` carries handedness, and quantized normals/tangents come back in -1..1 while quantized UVs and colours come back in 0..1, so the values are the same ones a surface shader sees whichever layout the mesh is in. ## shader.wgsl Write only `@compute fn main` (and any helpers / `#include`s). The names declared in `bindings.yaml` are in scope — `data`, `out_tex`, `params.scale`, … — with no `@group`/`@binding` lines of your own. ```wgsl // @domain: compute // Bindings are declared in bindings.yaml — do NOT write @group/@binding here. @compute @workgroup_size(64) fn main(@builtin(global_invocation_id) id: vec3<u32>) { if (id.x < arrayLength(&data)) { data[id.x] = data[id.x] * 2.0; } } ``` ## Dispatch & runtime Everything a compute shader does, it does through its reference: | method | what it does | |---|---| | `ref:dispatch(opts)` | `{ buffers, workgroups, params? }` — bind one buffer per declared storage binding, in order | | `ref:dispatchOnVertices(opts)` | `{ model, buffers?, workgroups, params? }` — bind the mesh `model` names to the `vertices` binding, `buffers` filling the rest | | `ref:dispatchEx(opts)` | `{ resources, workgroups }` — for shaders mixing textures, samplers and storage textures: one `{ kind, name }` per declared binding, in order | | `ref:setParam(prop, value)` | write one `params:` scalar uniform | | `ref:createBuffer(name, opts)` | `{ type, len, usage? }` — a GPU buffer owned by this shader | | `ref:buffer(name)` | the buffer this shader last made under `name` | | `ref:getBindings()` | the parsed `bindings.yaml` schema, read and parsed again only after the file is written | | `ref:getSource()` / `ref:setSource(wgsl)` | read / overwrite `shader.wgsl` | | `ref:compile()` | compile now instead of on the first dispatch | | `ref:status()` | what the engine did with this shader's dispatches, one record per target | The `params:` uniform is engine-owned and takes no entry in `buffers` or `resources`, and neither does a `vertices` binding's layout. `setParam` writes the uniform every later dispatch reads. A dispatch that passes `params = { name = value }` reads those values instead, in a copy of the uniform only it binds, with every name it leaves out at the value the shader holds — so one shader dispatched twice in a frame over different data reads different values each time: ```lua lift:dispatch({ buffers = { a }, workgroups = { 4 }, params = { height = 1.0 } }) lift:dispatch({ buffers = { b }, workgroups = { 4 }, params = { height = 3.0 } }) ``` A name the shader does not declare fails that dispatch, and the failure names what it does declare. ### `dispatchEx` resources Each `resources` entry fills the binding at its position. Its `kind` says what the resource is, in the words `bindings.yaml` uses, and the binding decides how it is read. A slot takes the kind it is declared with; a `texture2d` slot also reads a storage texture, a side of a history pair or the scene depth, and a `storage2d` slot also writes this frame's side of a history pair: | binding's `bindings.yaml` kind | `resources` entry | `name` | |---|---|---| | `buffer` | `{ kind = "buffer", name = buf }` | a buffer handle from `ref:createBuffer` | | `texture3d` | `{ kind = "texture3d", name = vol }` | a 3D texture handle from `ref:createTexture3D` | | `storage3d` | `{ kind = "storage3d", name = vol }` | a 3D texture handle from `ref:createTexture3D` made with `storage = true` | | `storage2d` | `{ kind = "storage2d", name = target }` | a storage texture handle from `ref:createStorageTexture2D` | | `storage2d` | `{ kind = "history_curr", name = history }` | a history pair from `ref:createTextureHistory`: this frame's side | | `texture2d` | `{ kind = "texture2d", name = tex }` | a handle `ref:copyBufferToTexture` returns, or a texture asset's guid once `texRef:handle()` has made it GPU-resident | | `texture2d` | `{ kind = "storage2d", name = target }` | a storage texture handle, read as a texture | | `texture2d` | `{ kind = "history_prev", name = history }` / `{ kind = "history_curr", name = history }` | a history pair: the previous frame's side or this frame's | | `texture2d` | `{ kind = "scene_depth" }` | none: the frame's scene depth | | `sampler` | `{ kind = "sampler", name = smp }` | a sampler handle from `ref:createSampler` | `texture3d` and `storage3d` name the same resource, a 3D texture, so either one fills either slot. The engine's own spellings name the same resources and keep working: `texture_3d` for `texture3d` and `storage3d`, `texture_2d` for `texture2d`, and `storage_2d` for `storage2d`. Any other `kind` is refused at the call, naming every kind it accepts. A `vertices` binding is filled by `ref:dispatchOnVertices`, and `texture_depth`, `textureCubeArray` and `acceleration_structure` bindings by a render feature's pass. ```lua local march = asset.resolve("volume_march", "computeShader") -- bindings.yaml: density (storage3d), out_tex (storage2d), U (buffer) march:dispatchEx({ resources = { { kind = "storage3d", name = density }, { kind = "storage2d", name = target }, { kind = "buffer", name = uniforms }, }, workgroups = { 8, 8, 8 }, }) ``` `workgroups` is `{ x, y, z }`, `{ x }`, or a bare `x`, and every dimension it leaves out is 1. A zero in any dimension is refused: the dispatch is dropped and recorded as a failure, so it surfaces in `ref:status()` and `compute.failing()` with the reason. A count derived from how much data there is passes through `math.max` first, so an empty set dispatches one workgroup rather than none: ```lua carve:dispatch({ buffers = { heights }, workgroups = { math.max(1, math.ceil(#points / 64)) } }) ``` ### Buffers and readback (the CPU side) A buffer is created through the shader that owns it and comes back as a handle. Pass the handle to a dispatch — including another shader's — to bind it there: ```lua local double = asset.resolve("@builtin::shaders.compute_double", "computeShader") -- `type` is the element the shader declares, `len` how many of them. local data = double:createBuffer("data", { type = "f32", len = 4, usage = { "readback" } }) data:write({ 1.0, 2.0, 3.0, 4.0 }) -- :writeU32 for integer words, :writeBytes for packed bytes double:dispatch({ buffers = { data }, workgroups = { 1, 1, 1 } }) local pending = data:read() -- starts an async GPU->CPU readback -- a frame or two later, drain it (nil until it has arrived): local out = pending:result() -- :resultU32 / :resultBytes read the same bytes another way data:destroy() -- free it when done ``` `type` is one of `"f32"`, `"vec3"`, `"vec4"`, `"quat"`, `"mat4"`, and `usage` adds `"readback"`, `"vertex"`, `"index"` or `"indirect"` on top of the storage a buffer always has. An `"indirect"` buffer holds a draw's arguments where the GPU reads them, so a shader that compacts geometry writes the drawn count into the same dispatch that produced it and a render feature's `kind = "draw"` pass submits exactly that — the `renderFeature` asset type's README documents that pass and the `storage.indirect` name it reads. The handle's key is the shader's guid plus the name you gave it and a serial, so two shaders that both want a buffer called `data` get one each — and so do two callers asking this same shader for one. `ref:buffer(name)` returns the last one made under that name, which is how two callers share one on purpose. A buffer that no shader computes — a vertex or index run that is only ever drawn — is allocated straight from the primitive instead: `substrate.createBuffer({ name = ..., type = ..., len = ..., kind = "gpu", usage = { "vertex" } })`. The first dispatch of a shader compiles it, so the very first readback can still reflect the un-run shader; dispatch again once it is compiled, or `ref:compile()` to pre-warm. This create/write/dispatch/readback loop is the CPU side of the resource model in `core/resource-model`. ### Is it compiled? `ref:compile()` queues the compile; the pipeline is built on the render side frames later, so that call cannot say whether it produced a program. `compute.programState(ref)` is where that lands: ```lua if compute.programState(carve) == "ready" then carve:dispatch(opts) end local state, reason = compute.programState(carve) if state == "failed" then print(reason) end ``` - `"absent"` — the engine holds nothing under this key and nothing is in flight: never asked for, or released. - `"pending"` — a compile was asked for and has not reached the device. A recompile of a resident program reads pending too, because what it produces is a different program from the one bound now — which is what makes this the precondition to wait on after an edit. - `"ready"` — compiled and resident, so a dispatch naming it binds it. - `"failed"` — the most recent compile produced no program, and the second return says why. A dispatch issued while the program is `"pending"` is held behind the compile and dropped if it never lands, so a readback taken before `"ready"` reports the buffer as it was. Wait for `"ready"`, not for a count of frames. ### Did it run? A dispatch is recorded into a command encoder frames after the call that asked for it returned, so that call's `true` means "queued", not "ran". `ref:status()` is where the outcome lands — one record per target, cumulative: ```lua for _, d in ipairs(carve:status()) do print(d.target, d.ok, d.dispatches, d.failures, d.lastError) end ``` `target` is the mesh guid for a `dispatchOnVertices` and the buffers it bound for a dispatch that writes only those, so two systems driving one shader over their own buffers read as the two passes they are. `ok` is the most recent outcome; `lastError` says why the last failure failed and is kept after a recovery. An empty result means nothing has dispatched this shader at all. `compute.failing()` answers the same question across every shader at once. A failing dispatch also posts a notice, so a pass that stops running reports itself rather than waiting to be found. ### Running one every frame A `kind = "compute"` substrate job queues one dispatch per frame: ```lua jobs.register({ phase = "main", executor = { kind = "compute", shader = asset.resolve("carve", "computeShader"), buffers = { "heights" }, workgroups = { 64 }, }, }) ``` ## Registration & hot reload Writing `shader.wgsl` or `bindings.yaml` fires the `.computeShader` assetType's `onChange`, which calls `__compute.compile` — that generates the binding declarations, expands `#include`s, naga-validates, and registers. This happens on create and on every edit, so saving recompiles with no world reload. A shader with invalid WGSL (a type mismatch, an undefined name) **fails to compile** with the reason logged (`[compute] Compile of '<name>' failed: …`) and is **not** registered. You don't need to register a shader before dispatching it — the first dispatch auto-compiles it on use. `ref:compile()` remains as an optional explicit pre-warm, only to avoid the one-frame first-dispatch warm-up in a latency-critical spot. ## What compute is holding, and why a key names nothing Every buffer, volume, storage target, history pair and sampler a dispatch can bind is filed under a key, and `compute.observe()` lists all of them with what each costs: ```lua local r = compute.observe() print(r.totals.count, r.totals.bytes) for _, res in ipairs(r.resources) do print(res.key, res.kind, res.bytes, res.owner.name) end ``` A row carries the key verbatim — the same string a dispatch names it by and a dispatch-failure message prints — plus its `kind`, its `bytes`, its extent and `format` where it has them, the `usage` bits it was created with, and the frame it was created on. `owner` is read back out of the key (`<shader-guid>::<name>#<serial>`), so a resource nothing has ever dispatched over still names the shader that made it. `compute.resources(ref)` narrows to one shader's — including the `<shader-guid>::__params` uniform the compile created for a `params:` block, which the shader owns like any other resource. `r.totals.bytes` is the `compute` figure of `renderer.gpuMemory()` — the two are read off the same registries, so a listing accounts for the category exactly. The same reading is served at `/zero/runtime/compute`. When a key names nothing, `compute.diagnose(key)` says which of the states the inventory distinguishes it is in, from the closed set `compute.absentReasons()` enumerates: | reason | what the inventory read | |---|---| | `resident` | a resource is filed under exactly this key; the row is on `resource` | | `replaced` | the owner holds the same name at a different serial — `current` names the live key | | `nameUnused` | the owner holds resources, none under this name | | `ownerAbsent` | no live resource carries this key's owner | | `unowned` | the key carries no owner segment, and nothing is filed under it | | `notPublished` | no renderer has published an inventory, so no key can be answered for | So a key out of `[compute] dispatch '…' failed` resolves against what exists, and a typo reports which part of it the engine could not match rather than coming back empty. A reason is the inventory at the moment the call read it. The reading is republished on a frame where a registry gained or lost an entry, so a resource created and diagnosed inside one script sees the reading from before it was made; wait a frame and ask again. ## Discovery - `asset.list("computeShader")` — every registered compute shader. - `asset.inspect("<name>")` — status, source, parse errors, this README. - `computeRef:getBindings()` — the parsed `bindings.yaml` schema (`{ bindings, params }`). - `compute.programState(ref)` — whether the program is compiled, still compiling, or failed, and why. - `compute.observe()` — every GPU resource compute holds, and what it costs. - `compute.diagnose(key)` — whether a key names a live resource, and why not. - `profiler.gpuFrame()` — what each dispatch of this shader spends on the GPU. A dispatch is timed under `compute.<identity>` — the same identity `asset.resolve` takes, so a timing row names the shader that produced it. ## Common pitfalls - **Compute shaders don't render** — verify output via a buffer readback or a debug pass, not a screenshot. - **`ref:compile()` returns before the program exists** — the pipeline is built on the render side; `compute.programState(ref)` says when it is `"ready"`, and a dispatch issued before that is dropped. - **A dispatch call returning true means queued** — ask `ref:status()` whether it ran. - **A destroyed buffer is refused where it is bound** — a handle outlives its storage, and `buf:alive()` reports which. - **Binding order is the contract** — `buffers` and `resources` bind in `bindings:` list order; reordering the list reorders the bindings. - **Don't write `@group`/`@binding` in `shader.wgsl`** — the engine owns them; re-declaring is the fastest way to a layout mismatch. - **Don't infer a mesh's vertex stride** — declare `kind: vertices` and use the accessors. A wrong stride writes positions correctly and normals into the wrong word, which renders as flat lighting rather than as broken geometry. - **Storage-texture format must be storage-compatible** — `r32f` / `rgba8` / `rgba16f` / `rgba32f`; others are rejected at compile. - **A key that comes back empty is not the same as one that failed** — `compute.diagnose(key)` separates a name never used from a generation that was replaced, and names the live key in the second case. ## Related types - `.shader` — render-domain shaders (surface, sky, post-process, screen) and the legacy `@domain: compute` fallback. ## In the Inspector Selecting a compute shader opens its **Compute** section: whether its program is ready (and why not), how many dispatches it has run and failed in this session (and the last error), tables of the bindings and params its `bindings.yaml` declares, and openers for `shader.wgsl` and `bindings.yaml`.

# guide A guide is a single-file documentation asset: a `<name>.guide/` folder whose body is `guide.md`, surfaced through the `guides` MCP tool, the `man` shell builtin and the Docs panel. Drop one anywhere under `/zero/source/`, a `.package` included, and it shows beside the built-in guides. ## In the Inspector Guide: the guide's word count, its headings, and **Open in Docs**, which opens it in the Docs panel.

# Importer (asset type) An importer is a Luau function that converts a source file (GLB, PNG, CSV, custom binary) into one or more derived assets registered in the engine. Importers extend the engine's file-format knowledge — adding a new importer means "the engine now understands `.foo` files". Importers run automatically on VFS writes that match their claimed extensions. ## When to use one - You're bringing a new file format into the engine and want to derive engine-native assets from it (meshes, textures, animations, data tables). - You want one VFS write to fan out into multiple derived assets (e.g. a GLB write derives mesh, texture, and animation assets). - You want format ingestion to be idempotent + hot-reloadable — importers re-run on source-file changes. If you only need to consume an existing engine asset type, you don't need an importer; just resolve it via `asset.resolve`. If you're building an authoring tool that emits assets, that's a `.tool`, not an importer. ## Where it lives - Source: `/zero/source/.../<name>.importer/` - Identity: `<name>` (the `.importer` suffix strips). - Folder shape: - `init.luau` (or `init.lua`) — exports `canImport` and `import`, and `extensions`: the file extensions it claims, lower-case without the dot. An importer that makes one asset beside its source also exports `produces`, the type of that asset (`"texture"` makes `<stem>.texture`). **Required.** - `README.md` — instance documentation. **Required.** - `.metadata` — agent-editable tags + free-form fields. **Required.** - `settings.luau` — the importer's own settings schema, declared with `Field` descriptors. Optional; an importer without one takes no settings. - `settings.yaml` — the importer's settings values. Optional. ## Settings An importer that takes settings declares them in its own `settings.luau` and receives the resolved values as `ctx.settings` in `import(ctx)`. `asset.schema(importerRef)` reports the schema, `asset.settings(importerRef)` the values, and `asset.setSettings(importerRef, patch)` changes them. What an importer stamps onto the assets it creates is typically one of these: the texture, SVG and model importers each declare `textureSettings = Field.settingsOf("@builtin::assetTypes.texture", NoSync)` — the texture type's own fields — and pass it to `asset.create` as `importerSettings`, the importer layer of each texture's settings. A change to an importer's settings counts as a change to what an import is made from, so an unchanged source is imported again after one. ## How to create one ```luau asset.create("importer", "<name>") -- Creates: /zero/source/<name>.importer/ -- init.luau (canImport / import stubs) -- README.md (instance README template) -- `folder` places it in a subfolder of /source instead of the root: asset.create("importer", "<name>", { folder = "importers" }) ``` ## How it operates 1. **Registration.** Writing `init.luau` into a `.importer/` folder indexes the importer with the importer registry. 2. **Dispatch.** When a file is written to VFS, the engine asks every registered importer `canImport(path, bytes)` (cheap predicate). The first importer that accepts dispatches to its `import` function. 3. **Derivation.** `import(path, bytes)` returns a list of `{ path, kind }` pairs naming the assets to derive. The engine instantiates each derived asset against its registered category. 4. **Idempotency.** `import` MUST be deterministic — importing the same source bytes twice produces identical derived assets. The importer pipeline uses content-hashing to short-circuit re-imports when bytes haven't changed. 5. **Hot reload.** Editing the importer reloads it; subsequent writes use the new logic. Already-imported assets stay imported with the old derivation until their source is imported again — writing the source path re-dispatches the importers, and `tools.use("importers", "run", <path>)` re-imports a target, many targets, or a whole folder on demand. 6. **Play.** A source written while play runs lands on the play shadow and imports there, so the source and its asset are kept or discarded together at play-exit. A source play holds where it stands (written before play, or by a durable write during it; `vfs.playHoldsInPlace`) cannot be moved into a new asset while play runs, so its import is owed: the job reads `owed` (reason `play-lock`), a notice names it, and it imports when play ends or pauses. `tools.use("importers", "queue")` lists owed imports and `importers.cancel` drops one. 7. **A lost live session.** While the live world session is lost and reconnecting (`vfs.sessionLost()`), every write to authored source is refused, so an import asked for then is owed rather than started, and one a drop catches part way is owed rather than failed, its job carrying the refused write as `error`; when the import had already moved its source into the asset it makes, it runs again on the source there, regenerating that asset around it. The job reads `owed` (reason `session-lost`), the notice `import owed until the live session re-establishes` names it, and it imports once the session is back, when the notice `import resumed: the live session re-established` names it and `import finished` (or `import failed`) settles it. A generation waiting on such an import completes with the asset it produces. Content this machine keeps to itself (`vfs.isSaveExcluded`: `/source/tmp/`, `/source/local/`) is written while the session is lost, so an import of a source there into an asset there runs at once. ## Re-importing over an asset an import made An import keeps its source inside the asset it makes: `art/plate.svg` becomes `art/plate.texture/` holding `plate.svg`. `import(ctx)` reads where to write from `ctx.output` (`<dir>/<stem>.<produces>` on a first import) and keeps the source with `ctx.retainSource(dest, opts?)`, which moves it to `dest` and returns where it is now. `import` returns `{ output, source }`, `source` being the file the asset keeps, which the asset's `import` provenance records as its source. - **The source written again beside the asset.** Writing `art/plate.svg` once more re-imports it into the same `plate.texture`: the new bytes replace the svg the texture retains (which keeps its guid), the written file leaves, and the asset is made again from the new bytes. Its `import` provenance counts the iteration on. - **The same bytes written again.** When the source's checksum and the importer's settings match the import that made the asset, no import runs: the asset already holds those bytes, so the written copy is folded into it, and `writeOutcome` answers `skipped` with reason `unchanged-content`, naming the asset and the file that holds the bytes. - **The asset itself.** `tools.use("importers", "run", <asset>)` re-imports a produced asset in place from the source it retains (`ctx.container` names it), so rewriting `plate.texture/plate.svg` and running it re-bakes the texture. A run that makes nothing says why: each skipped result carries a `reason` (`missing`, `no-source`, `no-claimant`, `unclaimed`). ## Discovery - `asset.list("importer")` — every registered importer. - `asset.inspect("<name>")` — claimed extensions, derivation conventions, source, this type README. - `ls /zero/source/<name>.importer` lists the files the folder holds. - `tools.use("importers", "list")` — every registered importer, with its name and identity. - `tools.use("importers", "explain", <path>)` — which importers claim a given source and which gate would block the import. ## Where a write's bytes end up An importer that claims a written source may move it into the asset it makes (the texture importer turns `shots/a.png` into `shots/a.texture/` holding `a.png`), so the path the write named stops holding the bytes. `require("@builtin::assetTypes.importer.shared").writeOutcome(path, sinceFrame, timeoutSecs)` answers where they went, for a write made on engine frame `sinceFrame` (`time.frameCount()` read before it): nil when no importer claims them, else the import's `state`, the `importer`, the asset made (`output`), the file the bytes are in now (`source`) and a `note` saying all of it. Every write that fires its side effects is offered to the importers by the same rule, whichever route made it, and the MCP `write_file` and `upload_file` answers carry this outcome as `import`. A quiet write (`{ quiet = true }`, and every `capture` save) is offered to no importer, and its bytes stay where they were written; `writeOutcome` answers `not-started` for one an importer would claim. A write whose import the running play session holds back (a fresh import moves its source into the asset it makes, and play holds a source that predates it, or that a durable write made canonical, where it stands) answers `owed` at once, with `reason` `play-lock`: the bytes stay at the written path, and the import runs when play ends or pauses. One the lost live session holds back answers `owed` with `reason` `session-lost`, and runs when the session re-establishes. `writeOutcomes(paths, sinceFrame, timeoutSecs)` answers a run of writes made together (a model, then its buffer and textures) once all of them have landed, within one wait: a folder `write_file` or `upload_file` lands every file without waiting on any import, then answers each this way. `write_file { awaitImport = false }` lands one file of such a run without waiting on its import. ## Authoring conventions - Keep `canImport` cheap — it's called for every VFS write. Match on extension or a quick header probe; defer real parsing to `import`. - Declare the extensions the importer claims as `M.extensions` and have `canImport` claim by that list. The Inspector lists them, and `require("@builtin::assetTypes.importer.shared").extensionsOf(importerRef)` answers them (lower-case, without the dot; nil for an importer that declares none). - Name derived assets predictably (e.g. `imported/<source-stem>/<asset-name>`). Predictable paths let callers reference derived assets directly. - Make `import` pure relative to its inputs. Don't read mutable globals or side-channel state. - Log derivations via `print` / the script log so authors can see what got produced from each source. ## Common pitfalls - **Expensive `canImport`.** It runs on every VFS write; an expensive predicate slows down the entire authoring loop. - **Non-idempotent `import`.** Re-importing produces different output → content hashing falsely caches stale results. Use only deterministic transforms. - **Conflicting importers.** Two importers claiming the same extension causes ambiguous dispatch. First-registered wins, but this is fragile — coordinate ownership of file extensions. - **Derived-asset suffixes.** Make sure derived paths end in the right `.<category>` suffix; otherwise they won't register as recognized assets. ## Related types - `.module` — for plain ingestion helpers that don't need the importer dispatch path. - `.tool` — for one-off conversions an agent invokes manually. - `.service` — if your importer needs persistent state (e.g. a cache); the service owns it, the importer reads from it. ## In the Inspector Imports: the extensions the importer claims (`M.extensions`), each file it imported in this session with the state its job reached, and **Find assets it produced**, which lists the assets in the world whose import provenance names this importer.

# renderFeature A **render feature** inserts custom GPU passes into the frame pipeline. The engine provides one generic capability — a per-frame render-pass queue and a set of pipeline-phase insertion points — and a feature is the authored Luau that decides what passes to enqueue, and where, every frame. ## Authoring contract A `.renderFeature` folder holds an `init.luau` that returns a feature module: ```luau return { setup = function(ctx) -- optional; runs once when the feature goes live -- A "fragment" pass's `shader` names a POST EFFECT, not a `.shader` -- guid directly — postprocess.add compiles the shader asset under -- that name once, here. Enqueueing a name never passed to -- postprocess.add runs the pass as a no-op every frame (the log -- reads "effect '<name>' not compiled — pass skipped"). postprocess.add("my_effect", asset.resolve("my.shader.guid", "shader"), { enabled = false }) end, render = function(ctx) -- required; runs every frame -- enqueue one or more passes. The spec shape depends on `kind` (see -- "Pass kinds" below): this is a "fragment" pass (full-screen shader over -- input textures); a mesh effect (outline, x-ray) is a "geometry" pass -- with `material` + `select` (see "Geometry passes"). ctx.enqueue { phase = "afterTransparent", shader = "my_effect", inputs = { src = "@scene.color" }, output = myRenderTargetGuid, } end, teardown = function(ctx) end, -- optional; runs once when the feature is destroyed } ``` Optional `materials/` and `shaders/` subfolders hold shader source the feature constructs GPU programs from. A feature constructs its GPU programs (it does not reference `.material` surface assets). ## `ctx` Built fresh each frame and passed to `setup` / `render` / `teardown`: | Field | Meaning | |-------|---------| | `ctx.enqueue(spec)` | Queue one render pass for the current frame. | | `ctx.viewport` | `{ w, h }` — the **main viewport's** size in pixels. | | `ctx.time` | Seconds since boot. | | `ctx.frame` | Monotonic frame index. | | `ctx.renderPath` | `"forward"` / `"deferred"` / … — branch passes on the active path. | | `ctx.settings` | The feature's knobs as they stand this frame — see "Settings". | ## Settings A feature declares the knobs it takes as `settings`, each a Field descriptor with a default, the way a component declares its public fields: ```luau local Fields = require("@builtin::modules.field") local Field, NoSync = Fields.Field, Fields.NoSync return { settings = { intensity = Field.number(1.0, NoSync, "How much of the effect reaches the image."), quality = Field.enum({ "low", "high" }, "high", NoSync, "How many samples the pass spends."), }, render = function(ctx) local s = ctx.settings -- { intensity = 1.0, quality = "high" }, or what a caller set ... end, } ``` The engine holds one settings table per live feature — every declared default, overridden by what it was handed — and passes it as `ctx.settings` on every `setup` / `render` / `teardown`. A feature reads its knobs there and nowhere else, so any caller tunes any feature by name: ```luau local h = renderer.feature.create(ref, { settings = { intensity = 0.5 } }) renderer.feature.configure(h, { quality = "low" }) -- lays over the rest renderer.feature.settings(h) -- { intensity = 0.5, quality = "low" } renderer.feature.find(ref) -- the live handle, or nil renderer.feature.schema(ref) -- the declared knobs, unrun ``` A knob the feature does not declare, or a value of the wrong shape, is refused by name. A scene authors its features on one `RenderFeatures` entity — a list of features, each with its settings — which adopts a feature already live rather than stacking a second, and takes every one down on disable and destroy. `render` runs **once per frame**; the passes it enqueues run **once per render target** drawn that frame — the live viewport, every offscreen capture, every RTT camera — each at its own resolution. So `ctx.viewport` is the main viewport's size, not the size of the target a pass ends up drawing into. Size a full-screen scratch RT with `screen = true` and let the engine track the drawn target (see "Screen-sized targets") rather than deriving it from `ctx.viewport`. ## Screen-sized targets `renderer.texture.create { screen = true }` marks a feature RT as full-screen: before each render target's feature passes, the engine resizes it to that target's resolution. Create it once — the engine owns its size from then on: ```luau myRt = renderer.texture.create { width = ctx.viewport.w, height = ctx.viewport.h, -- initial size only screen = true, format = "rgba16f", name = "my_effect", } ``` A target that holds a picture of the scene is `rgba16f`, which is what the scene target itself carries: linear light with headroom above the display's white and no 256-step floor under a dark gradient. Omitting `format` on a screen-sized target asks for the same thing. `format = "rgba8"` is for a mask, a weight or an occlusion term, where 0..1 is the whole range — hand one of those back over `@scene.color` and the frame loses every highlight and bands every gradient, and the engine says so in the log. `screenScale` (default `1.0`) takes a fraction of the drawn target's resolution, for a scratch that runs below composite resolution (a half-res blur passes `0.5`). A screen-sized RT is what makes a full-screen effect composite correctly at any resolution. Deriving the size from `ctx.viewport` instead pins the RT to the main viewport, and every differently-sized render target then composites a mismatched scratch — the scene scaled into a corner with the rest black. `screenSpace` names WHICH image a screen-sized RT follows: | value | what it tracks | |---|---| | `"scene"` (default) | the scene render target, at whatever resolution the render scale asked for. Resized before EVERY render target's feature phases — the main frame and each offscreen one — and cleared to transparent before an offscreen render, so one sink never reads what another left behind | | `"composite"` | the image the post-scene phases draw into. Resized on the presented frame only, and never cleared for an offscreen render, since none of those phases run there | The two carry the same size until the render scale drops below 1. From there, while the renderer presents the viewport itself, the scene reaches the display's resolution through an upscale that runs before `afterScenePost` — so a composite-space RT is the display's size while a scene-space one is the smaller rasterized frame. Under a UI viewport panel the panel draws the scene target at its own rect, nothing upscales beforehand, and both spaces carry the scene's size. A scene-space target is scratch: whatever a pass does not rewrite this render is transparent black. A composite-space target survives an offscreen render untouched, but it follows the presented frame alone. ## History targets An accumulation across frames (temporal antialiasing, a temporal denoiser, any pass that reads what it wrote last time) lives in a history target: ```luau history = renderer.texture.create { width = ctx.viewport.w, height = ctx.viewport.h, -- initial size only history = true, format = "rgba16f", name = "my_accum", } -- in render(ctx): ctx.enqueue { kind = "compute", program = myResolve, inputs = { scene_color = "@scene.color", accum_previous = history.previous.guid }, buffers = { accum_state = history.state }, storage = { accum_out = history.guid }, phase = "afterLighting", } ``` The passes a feature enqueues run once for every view the frame draws (the viewport, each render-to-texture camera, each capture), so an accumulation has to be one per view. A history target keeps one pair of images per view, sized to that view: for the length of a view's render `history.guid` names the image the render writes and `history.previous.guid` the image the same view's previous render wrote. Another view never resizes, clears or overwrites them, and a view's images are freed when it stops drawing. It follows `screenSpace` and `screenScale` like any screen-sized target; a composite-space one is kept for the viewport, the one view that draws the post-scene phases. Before each render of a view the engine writes `history.state`, one `vec4<f32>` a pass binds through `buffers`: | field | holds | |---|---| | `x` | 1 when `previous` holds this view's previous image; 0 on the view's first render, its first after a camera cut or a change of size, or after a render of the view that ran without the target, and `previous` then reads transparent black | | `y` | how many renders in a row have carried the history | | `zw` | this render's sub-pixel projection offset, in pixels (`renderer.setProjectionJitter`) | `renderer.setProjectionJitter` is the matching per-view sample sequence: every view walks it by its own count of renders, so the sequence a temporal pass expects is the one each view's renders sampled. ## Pass phases `enqueue` runs a pass at a named pipeline boundary — a real seam between the scene's discrete passes — ordered within a phase by `order`. Each phase is the point where the scene colour holds a specific, nameable state: | Phase | Scene colour there | |-------|--------------------| | `afterDepthPrepass` | the visibility raster written: `@scene.depth` holds this frame's complete opaque depth, and `@scene.motion` its opaque motion, before anything shades. Scene colour is not yet drawn; `@scene.colorHistory` holds the previous frame's lit colour, and the deferred g-buffer the previous frame's surfaces (`readsPrevious`) | | `afterShadows` | shadow maps populated; scene colour/depth not yet drawn | | `afterSky` | sky only; depth empty | | `afterOpaque` | opaque drawn, before transparents. `@scene.depth` (and the deferred G-buffer) are populated. Forward: lit opaque. Deferred: opaque is in the G-buffer, so the colour is still sky only until the composite (use `afterOpaqueLit`) | | `afterOpaqueLit` | the lit opaque scene, still before transparents, on **both** paths — the forward opaque draw, or the deferred lighting composite. The background a refractive or transmissive surface samples: the one phase whose colour holds what is behind such a surface without holding the surface itself | | `afterTransparent` | transparents drawn, before the OIT composite | | `afterLighting` | deferred-lighting + OIT composite done — fully lit, final geometry on **both** paths | | `afterScenePost` | scene-layer post (DoF/MB/…) done, before UI | | `afterUI` | scene + UI composited | | `afterPost` | all-layer post done — the final image, before present | The last three are the post-scene phases, and they work on the COMPOSITED image: `@scene.color` there is the target the UI lands on, not the scene render target. `@scene.depth` and `@scene.motion` still bind — they are the scene's own buffers for the frame just drawn, which is what lets a pass at these phases reproject the composite it is handed — but they keep the SCENE's resolution. Whenever the render scale is below 1 on the presented frame, that is smaller than `@scene.color`, so a pass here maps between the two by uv rather than by texel index, and takes its own output size from the target it writes. The `@scene.*` inputs (`@scene.color`, `@scene.depth`, …) bind the live frame buffers; the deferred g-buffer channels (`@scene.normal`, `@scene.shaded`, …) resolve only on the deferred path. Before the pass that shades has drawn — at `afterDepthPrepass` — a g-buffer channel holds the previous frame's surfaces, and a pass reading it there names the slot in `readsPrevious` (`readsPrevious = { "scene_normal" }`); one that does not is held back rather than handed a stale normal. `@scene.colorHistory` is the lit colour the target's previous render ended with, kept by the engine per render target and never cleared by a frame: what a pass tracing at `afterDepthPrepass`, against this frame's depth, takes its light from — reprojected through the previous view-projection `@frame.camera` carries. A capture drawn once, or a camera drawn again after a pause, is drawn lit first when a pass reads the history, so its first frame holds one. Every `@scene.*` channel, what it holds, its format, the phase it is first valid from, and the earlier spellings still bound to it: | Name | Holds | Format | Valid from | Aliases | |---|---|---|---|---| | `@scene.color` | the scene's lit colour; the display composite at a post-scene phase | the scene target's format at a scene phase, the display composite's at a post one | every phase | — | | `@scene.depth` | opaque depth, reversed (1.0 near, 0.0 far/empty sky) | `Depth32Float` — binds through a `texture_depth` compute slot, not `texture2d` | `afterDepthPrepass` on | — | | `@scene.motion` | screen-space NDC velocity since the previous frame | `Rg16Float` | `afterOpaque` on | — | | `@scene.normal` | world-space normal + the per-pixel surface-flag field | `Rgb10a2Unorm` | deferred path, `afterOpaque` on (previous frame's before — declare `readsPrevious`) | — | | `@scene.shaded` | what a surface SENT — its shaded colour, linear and scene-referred | `Rgba16Float` | deferred path, `afterOpaque` on (previous frame's before) | `@scene.albedo` | | `@scene.material` | what a surface IS — base colour (RGB565) + metallic + perceptual roughness | `Rgba8Uint` | deferred path, `afterOpaque` on (previous frame's before) | `@scene.mr` | | `@scene.emissive` | radiance a pass stamped on the surface, added at the composite | `Rgba16Float` | deferred path, `afterOpaque` on (previous frame's before) | — | | `@scene.ao` | scene occlusion a feature wrote, read back by shading | `Rgba8Unorm` | every scene phase (cleared at the top of the scene render) | — | | `@scene.reflection` | traced reflection radiance (`.rgb`) + confidence (`.a`) | `Rgba16Float` | every scene phase (cleared at the top of the scene render) | — | | `@scene.colorHistory` | the lit colour the target's previous render ended with | `Rgba16Float` | every phase | — | | `@scene.lights` | the punctual-light storage buffer: 64 light entries of twelve vec4s each (the light, then its punctual shadow record), followed by the virtual shadow map's region. A loop runs over the first `@frame.camera[4].w` entries, the lights in use; the buffer's length counts the shadow region too, so `arrayLength` does not size a light loop | storage buffer | every phase | — | | `@scene.environment` | the reflection-probe cube array | texture cube array | scene phases only | — | | `@scene.environmentSampler` | the filtering sampler for `@scene.environment` | sampler | scene phases only | — | | `@scene.reflectionProbes` | the environment buffer: the probe blend data naming which probe reaches a point, then the irradiance light-probe volumes | storage buffer | scene phases only | — | `docs/topics/rendering.guide/guide.md` and `docs/topics/raytracing.guide/guide.md` both point here rather than repeating this table. **Two g-buffer channels describe the surface, and they answer different questions.** `@scene.shaded` is what a surface SENT — the colour the geometry pass shaded it to, linear and scene-referred, so a surface lit above the display's peak reads above 1 there. `@scene.material` is what the surface IS: the base colour it reflects, its metallic and its perceptual roughness, each in [0, 1] and each independent of how brightly the surface happens to be lit. A pass that multiplies incoming light by a reflectance — a bounce multiplier, a Fresnel tint, an ambient floor — reads `@scene.material`; a pass that filters the picture reads `@scene.shaded`. `@scene.material` is an `Rgba8Uint` channel packing all three into 32 bits, so its slot declares `sample: uint` and the module unpacks it: ```yaml # bindings.yaml - { name: scene_material, kind: texture2d, sample: uint } ``` ```wgsl #include "@builtin::shaderModules.gbuffer_material" let m = zero_gbuffer_material(scene_material, coord); let bounce = incoming * m.base_color; ``` The earlier spellings still bind: `@scene.albedo` resolves to `@scene.shaded` (the SHADED lit colour) and `@scene.mr` resolves to `@scene.material` (the material channel), so a feature written against either keeps resolving. The engine logs the substitution once per spelling, naming the canonical one to write going forward. `@scene.motion` carries screen-space velocity: the NDC displacement each pixel's surface underwent since the previous frame, so a still surface under a still camera reads `(0, 0)`. Opaque geometry contributes it on both render paths: the visibility raster writes it for every surface it records, and an opaque surface whose depth state the raster cannot express writes its own where it draws. A moving entity, a moving camera and an animated pose all reach it in the viewport; transparents, particles and sky carry none. A surface a material's `vertex()` hook displaces carries the displacement's own velocity. A render target's own camera and a capture keep no view history, so velocity there is the entity's motion with no camera term. `@frame.camera` is a reserved **buffer** input carrying the camera CURRENTLY being drawn: 10×`vec4<f32>` = the inverse view-projection (columns `[0..4)`, column-major) + camera world position (`[4].xyz`) + the previous render's view-projection (columns `[5..9)`), which projects a point into the picture that render left. A compute pass declares a `buffer` binding and binds it via `inputs = { name = "@frame.camera" }`, then reconstructs a pixel's world position from `@scene.depth`: ```wgsl let inv_vp = mat4x4<f32>(cam[0], cam[1], cam[2], cam[3]); // ndc.z is the `@scene.depth` sample itself. Scene depth is reversed — 1.0 is // the near plane, 0.0 the far plane and empty sky — and this matrix inverts // that same range, so a sample goes in unmodified. let wh = inv_vp * vec4<f32>(ndc, 1.0); let world_pos = wh.xyz / wh.w; ``` Like `@scene.*`, the engine resolves it **per render target** — so the same pass reconstructs against the live viewport's camera on screen and against an offscreen capture's / RTT camera off screen. (Pack a camera matrix into your own buffer instead and the offscreen render reconstructs against the wrong camera.) > Render-feature passes run in the main viewport render **and** in offscreen > captures / render-to-texture, each against its own camera. The `capture` tool's > `source = "position"` / `"entity"` therefore run the pass queue too; using > `@frame.camera` is what keeps those captures matching the live viewport. ### The environment A compute pass reads the scene's captured surroundings — the reflection-probe cube array, its sampler and the environment buffer (the probe blend data, then the irradiance light-probe volumes) — through three reserved inputs, bound to bindings of the shapes group(0) binds them in: ```yaml bindings: - { name: environment_map, kind: textureCubeArray } - { name: environment_sampler, kind: sampler } - { name: zero_environment, kind: buffer, access: read, element: ZeroEnvironment, array: false } ``` ```luau inputs = { environment_map = "@scene.environment", environment_sampler = "@scene.environmentSampler", zero_environment = "@scene.reflectionProbes", } ``` Named like that, `#include "zero::environment"` in the shader resolves the environment at a world point along a direction — `zpbr_sample_environment(world_pos, dir, lod)`, the probes whose radius reaches the point blended by rank and proximity, each through its proxy volume, and the sky under them — exactly as a material's environment term does, so a reflection a pass draws and one a material draws agree about the room they are in. The three bind at the scene phases; a post or UI phase carries none. A `textureCubeArray` binding takes `@scene.environment` alone, and a fragment pass has no slot of that shape. ## Pass kinds `enqueue` runs one of three kinds (the `kind` field; default `"fragment"`): - **`"fragment"`** — a full-screen shader over input textures. The program is a GPU **material** (or `shader`) the feature constructed; `inputs` bind live textures (`@scene.*` or RT guids), `output` is the target RT (omit = the phase's default scene target). Post-process, blur, edge effects. **Writing named G-buffer channels (`outputs`).** Instead of the single `output` RT, a fragment pass may declare an `outputs` array — one or more targets, each a named engine channel (`@scene.shaded` / `@scene.normal` / `@scene.emissive` / `@scene.color`) or an RT guid — each with its own blend mode: ```luau ctx.enqueue { kind = "fragment", shader = myShaderKey, outputs = { { target = "@scene.shaded", blend = "under" }, -- premultiplied dst-over-src { target = "@scene.emissive", blend = "additive" }, -- one fullscreen pass, two targets (MRT) }, phase = "afterOpaque", -- deferred: G-buffer filled, before the lighting composite } ``` Blend modes: `"replace"` (overwrite), `"alpha"` (source-over), `"additive"` (sum), `"under"` (premultiplied dst-over-src — accumulates coverage-weighted contributions). A `@scene.*` channel resolves only where it exists (the deferred G-buffer at a pre-composite phase); a pass whose targets are all unavailable at its phase is skipped (logged). `output` + `blend = true` remains the shorthand for a single alpha-blended target. The shader's returned colour is written to every target in the set. A fragment shader returns `vec4<f32>`, so the channels it can target are the float ones. `@scene.material` (and its earlier spelling `@scene.mr`) is `Rgba8Uint`, which binds as `uint` — naming it as an output is refused by name, with the channel's format and class in the log, and the rest of the frame draws. A fragment pass's `inputs` reads through the same fixed float slot, so naming `@scene.material` there is refused the same way; a compute pass reads that channel through a `sample: uint` slot instead. **What a pass with no `inputs` samples.** A fragment pass that declares no `inputs` reads `@scene.color` — the shape of an effect that filters the scene image. A pass that names `@scene.color` among its own `outputs` samples a 1x1 transparent texel instead, because the scene colour is that pass's render target and one texture is a colour attachment or a sampled resource in a render pass, never both. Declare `inputs = { src = "@scene.color" }` and write an RT to read the scene image and write somewhere else in one pass. - **`"compute"`** — a workgroup dispatch (`shader`, `storage`, `buffers`, `dispatch`; see "Compute passes"). - **`"geometry"`** — draw scene geometry with a constructed pipeline. The program is a **material** the feature built (`renderer.material.create`); the engine draws the scene's render-objects with that material's pipeline + uniforms. - **`"draw"`** — rasterize the feature's OWN geometry: `geometry` names a buffer created with `usage = { "vertex" }`, `storage.index` one created with `usage = { "index" }`, and `select` the one entity whose transform and material the draw lands through. `params[0]` is how many indices to submit — or name a buffer under `storage.indirect` (created with `usage = { "indirect" }`, holding a `DrawIndexedIndirectArgs` block: index count, instance count, first index, base vertex, first instance) and the count comes from the GPU instead, so a compute pass that compacts geometry pays only for what it kept and nothing waits on a read-back. `first_instance` must be 0; the engine resolves the entity's per-instance data another way, because a non-zero one needs a device feature WebGPU does not guarantee. - **`"splat"`** — rasterize a compute buffer of particle records as instanced camera-facing billboards into feature RTs (see "Splat passes"). ### Geometry passes A geometry pass re-draws the scene's meshes with a material you supply — the generic primitive for mesh-space effects (the engine names none of them): ```luau ctx.enqueue { kind = "geometry", material = myMaterial, -- the HANDLE renderer.material.create -- returned (or its registry-key string) geometry = "@geometry.opaque", -- which drawables: "@geometry.opaque" | -- "@geometry.transparent" | "@geometry.all" select = { entityA, entityC }, -- OPTIONAL: restrict to these entities -- (entity refs); omit = every matching mesh phase = "afterLighting", -- afterLighting = lit scene + valid depth } ``` `select` restricts the draw to the given entities — the per-instance signal that turns a whole-scene pass into a chosen-meshes effect. The **material** decides how those meshes are drawn; you control three things: - **Pipeline render-state.** `renderer.material.create` takes a `render` block — `cull` (`back`/`front`/`none`), `depthWrite` (bool), `blend` (`opaque`/`alpha`/`additive`) — so the pass can rasterize differently from the surface pass: ```luau local myMaterial = renderer.material.create({ shader = myShaderKey, -- a .shader you authored (identity, not guid) render = { cull = "none", depthWrite = false }, properties = { outline_color = { 1, 0.4, 0, 1 } }, -- seed the shader's uniforms }, "my_pass_mat") -- pass `myMaterial` (or "my_pass_mat") as the pass `material` ``` - **The shader.** A geometry `.shader` is a surface shader: author `fn vertex(v: VertexData) -> VertexData` to move vertices (object-local `v.position` / `v.normal`) and `fn fragment(input: FragmentData) -> vec4<f32>` for the colour. - **Its properties.** A shader reads ONLY the uniforms IT declares in `properties.yaml`, via `material.<yourProp>`. Referencing a field it didn't declare (e.g. a StandardMaterial `material.base_color`) fails to compile and the mesh renders **magenta** — declare the property or use a constant. What effect those three produce (outline, x-ray, wireframe, highlight, …) is yours to design — the engine just draws the selected geometry with your pipeline. ### Compute passes A compute pass dispatches a `.shader` compute program over a workgroup grid: ```luau ctx.enqueue { kind = "compute", shader = myComputeShaderKey, storage = { outTex = myStorageRtGuid }, -- slot -> storage-texture RT guid buffers = { particles = "my_particle_buf" },-- slot -> named compute buffer inputs = { camera = "@frame.camera" }, -- slot -> @frame.*/@scene.* buffer dispatch = { x = 64, y = 1, z = 1 }, } ``` **`dispatch` is optional.** Omit it and the pass is dispatched over what it WRITES: a grid covering its first `storage` texture, at the `@workgroup_size` the shader's entry point declares. With a screen-sized storage target that grid tracks the render target being drawn, so a screen-space effect covers the live viewport, an offscreen capture and an RTT camera without the feature knowing any of their resolutions: ```luau ctx.enqueue { kind = "compute", shader = myComputeShaderKey, inputs = { scene_color = "@scene.color", cam = "@frame.camera" }, storage = { outTex = myScreenSizedRt.guid }, -- screen = true phase = "afterLighting", } ``` State an explicit `dispatch` for a grid that is not one thread per output texel (a reduction, a particle sim), or when the shader's `@workgroup_size` is an override/const expression the engine cannot read. Each WGSL storage-buffer binding in the shader resolves its bound slot in this order: an explicit `buffers` mapping (the handle a buffer's owner holds, or a name), then `inputs` (`@frame.camera` / `@scene.lights` / `@scene.reflectionProbes`), then — when a slot has neither — a buffer whose engine-side name equals the slot's own name. `buffers` is the seam a pass binds through: the buffers a shader works over are the ones handed to it, whatever each is called. `storage` binds storage TEXTURES (RT guids) the same way, one map per resource kind. ### Splat passes A splat pass draws every record of an f32 buffer as a camera-facing billboard — the raster primitive behind screen-space particle surfaces (the `fluidSurface` system is built entirely on it): ```luau ctx.enqueue { kind = "splat", phase = "afterLighting", output = depthRt.guid, -- or `outputs` for the thickness+color MRT splat = { buffer = "my_particles", -- compute-buffer name stride = 12, -- floats per record posOffset = 0, -- float offset of the position count = 200000, -- records to draw radius = 0.3, -- world particle radius channel = "depth", -- "depth" | "thickness" | "sprite" | "gaussian" velOffset = 4, -- optional: velocity → screen-space stretch stretch = 0.04, -- seconds of motion elongating the billboard colorOffset = 8, -- optional (thickness): + per-particle RGB → MRT weightOffset = 7, -- optional (sprite): per-particle intensity minY = -60.0, -- records below this y collapse (parked/dead) maxRadiusPx = 120, -- optional: cap on the projected pixel radius }, } ``` The `"gaussian"` channel draws the engine's packed splat records — what `renderer.splat.components` decodes a `.spz` or `.ply` capture into — as elliptical-weighted-average Gaussians, and takes fields of its own: ```luau splat = { buffer = "cloud_records", stride = 6, -- the packed record is six u32 words posOffset = 0, count = 1200000, radius = 1.0, -- multiplies each splat's stored scale channel = "gaussian", orderBuffer = "cloud_order", -- back-to-front permutation; "over" needs one shBuffer = "cloud_sh", -- optional: quantized higher-order SH shStride = 24, -- u32 words of SH per record shDegree = 3, transform = model, -- object -> world, column-major, 16 floats prevTransform = prevModel, -- where `transform` stood last frame opacityScale = 1.0, } ``` `transform` places the cloud: each Gaussian's covariance is conjugated by its upper-left 3×3, so a non-uniform scale shears the cloud the way it shears a mesh. `prevTransform` plus a second entry in `outputs` makes the channel write an MRT pair — the colour into the first target, and the screen-space displacement each splat underwent between the two placements into the second, premultiplied by the same alpha the colour composites under. That second target therefore resolves to `(coverage-weighted mean displacement, coverage)`, which is what folds a moving cloud into `@scene.motion`. Naming one target, or leaving `prevTransform` out, renders the colour alone. The other three channels write different fragments: `"depth"` writes nearest-surface linear view depth (R32Float, hardware-depth-tested on the target's own depth buffer), `"thickness"` accumulates additive optical thickness (with `colorOffset` and a two-entry `outputs` list it also accumulates thickness-weighted color into the second target), and `"sprite"` accumulates a soft additive disc scaled by `weightOffset`. Every channel discards fragments behind opaque scene geometry by sampling the scene depth at the fragment's own NDC. `maxRadiusPx` scales down billboards whose projected major radius would exceed that many pixels, so particles crossing the near plane stay droplet-sized instead of filling the screen. Targets clear at the pass; `blend = true` on the spec loads them instead to accumulate over a prior splat pass. `renderer.feature.create(featureRef)` instantiates a feature and returns a live handle (`featureRef` may be an `AssetRef<renderFeature>` or its identity string). The handle appears under `/zero/runtime/generated_features/<guid>/`. Re-creating the SAME feature asset replaces the live one (hot-reload after editing `init.luau`) rather than stacking a duplicate. - `renderer.feature.list()` → `{ {guid, identity} }` for every live feature. - `renderer.feature.destroy(handleOrGuid)` (or `renderer.destroy(handle)`) tears one down — by handle OR by guid string, so you can stop a feature even after losing its handle (e.g. across separate `execute` calls). ## In the Inspector Selecting a render feature opens its **Feature** section: whether the feature runs in this session, an Enable or Disable action, and its settings as typed controls. While it runs, the controls hold its live values and an edit reconfigures the running feature (`renderer.feature.configure` checks it against the schema first). Those edits are session state: they change the running feature only, they are not undoable (Ctrl+Z never reaches them), and Disable or a restart ends them; the feature's declared defaults are what the next Enable starts from. While it does not run, the controls show the declared defaults, read-only, and a write reaching it then is refused. A feature enabled, disabled or reconfigured from a script shows in the section as it happens. A feature that declares no settings says so.

# Shader module (asset type) A `.shaderModule` is a block of WGSL that other shaders **include** rather than copy — a lighting model, a set of colour-space helpers, one package's shared prelude. It declares no entry point and is never dispatched or drawn on its own account; it exists to be expanded into the shaders that name it. ## Folder shape ``` <name>.shaderModule/ module.wgsl # the WGSL — helpers, structs, constants (required) README.md # what THIS module provides (required) ``` Each file carries a committed `.meta` sidecar pinning its stable guid, and the folder carries one of its own. ## Including one A shader names a module by identity, and the identity is what records the dependency — so a shader that is packed, pulled or published carries the modules it includes with it: ```wgsl #include "~.colour" // a module in the same package #include "@builtin::shaderModules.pbr_shading" // one from the builtin library ``` A missing module is a hard error at compile: a shader that names an interface which no longer exists fails loudly rather than compiling against a stale copy. Each module is expanded at most once per shader, so two shaders that both include the same module — or a module that includes another — produce one copy, not a duplicate-definition error. ## Writing one Write only what other shaders need in scope: functions, structs, constants. A module may reference a binding its includers declare, which is how a package's prelude can read that package's uniform block. ```wgsl const LUMA: vec3<f32> = vec3<f32>(0.2126, 0.7152, 0.0722); fn srgb_to_linear(c: vec3<f32>) -> vec3<f32> { let lo = c / 12.92; let hi = pow(max(c + vec3<f32>(0.055), vec3<f32>(0.0)) / 1.055, vec3<f32>(2.4)); return select(hi, lo, c <= vec3<f32>(0.04045)); } ``` ## Answering to another name A module declares extra names it answers to with `@alias` lines in the header of `module.wgsl`: ```wgsl // @alias: colour_helpers // @alias: @mypack::shaderModules.colour ``` Each becomes an alias on the asset, so the name resolves everywhere an identity does — `asset.resolve`, an `#include`, and the publish gate alike. That is what lets a module be renamed without breaking the shaders already written against its old name: content persists the source it was authored with, so a name that stops resolving is a world that stops publishing. Only the header is read — the scan stops at the first line that is not a comment, so a directive cannot hide in the module body. A name another asset already holds is refused and reported, since an alias extends the identity namespace rather than taking a name out of another asset's hands. ## Registration & hot reload Writing `module.wgsl` registers the module under every name the asset answers to — its guid, its identity, and any alias — so a shader includes it by whichever name it holds. Saving re-registers it and the shaders that include it recompile, with no world reload. ## Discovery - `asset.list("shaderModule")` — every registered module. - `moduleRef:getSource()` — the module's WGSL. ## Related types - `.shader` — render-domain shaders (surface, sky, post-process, screen). - `.computeShader` — compute programs. ## In the Inspector Selecting a shader module opens its **Module** section: how many lines it holds, the aliases a shader `#include`s it by, and an opener for `module.wgsl`.

# Preset (asset type) A preset is a named set of typed values for one declarer's settings schema: the texture assetType's, an importer's, a component's, any asset that declares its own settings. Capture the settings something has now, then apply them to anything the same declarer governs, instead of setting the same values by hand each time. ## When to use one - The same settings recur across many assets or entities (e.g. "pixel art": every sprite texture with `filter = nearest` and no mipmaps; a tuned locomotion set for every character of one rig). - Designers swap configurations without editing code. - A configuration needs a stable name other content can refer to. If you're capturing a whole entity assembly, use a `.bundle`. If you're factoring shared logic (not configuration), use a `.module`. ## Where it lives - Source: `/zero/source/.../<Name>.preset/` - Identity: `<Name>` (the `.preset` suffix strips). - Folder shape: - `preset.yaml` — the target and the values. **Required.** - `README.md` — instance documentation. **Required.** - `.metadata` — agent-editable tags + free-form fields (`title`, `description`). **Required.** ## The document ```yaml target: { __ref: "8cc28fc8-ed63-48ee-83a6-15f4192e27e1", type: assetType, name: texture } values: filter: nearest generateMipmaps: false ``` - `target` names the declarer by a guid envelope: `__ref` is its guid; `type` and the plain `name` tell a person reading the file what it points at. - `values` holds only the fields this preset sets. Fields it leaves out keep whatever the asset or component has. A component preset has the same shape, its target the component: ```yaml target: { __ref: "18dfece3-fe1b-4630-b7ad-0996f5043a2e", type: component, name: Locomotion } values: walkSpeed: 1.4 runSpeed: 2.5 ``` ## How to create one ```luau -- Capture what an asset has now: local p = preset.capture("hero_sprite", "pixel_art") -- Capture a component's live values on an entity: local tint = preset.capture({ entity = id, component = "Model" }, "red_tint") -- Both land at /zero/source/presets/<name>.preset/ (an absolute `.preset` -- path lands where it says; `{ overwrite = true }` re-authors an existing -- preset, keeping its guid). -- Or author it directly: asset.create("preset", "pixel_art", { target = "@builtin::assetTypes.texture", -- an identity or a guid values = { filter = "nearest" }, }) ``` Each checks every value against the target's schema and refuses the whole write on any violation, naming each one. ## How to use one - `preset.values(p)` returns `(values, report)`. Every value is checked against the target's current schema: valid fields are returned, and each invalid one is listed by path in `report.rejected`. - `preset.apply(p, into)` writes the valid values into an asset the target governs, or into a component on an entity named `{ entity = <id>, component = <name> }` whose component is the target. A struct group merges field by field; any other field is written whole. Applying to something another declarer governs raises, naming both. - `preset.setValues(p, values)` replaces the whole values table, checked first. Inside an undo operation the write is recorded into it, so undoing the operation puts back what `preset.yaml` held. - A preset is a snapshot, not a live binding: applying it writes values once. ```luau preset.apply("pixel_art", "enemy_sprite") preset.apply("synty", { entity = id, component = "Locomotion" }) local values, report = preset.values("synty") ``` ## Defaults for new things — the Preset Manager A world's Preset Manager rows make a preset the starting point for what is created from then on: ```luau local p = preset.capture("hero_sprite", "pixel_art") -- or asset.create("preset", …) preset.addDefault(p, { filter = "sprites/**" }) -- new things only ``` Each row names a declarer, a preset for it, and a filter. For a declarer, the first row whose filter matches the new thing's path applies. The rows live in `.world_settings` as `[[presets]]` blocks, sync with the world, and are edited in the Settings panel's **Preset Manager** section. **The layers a new thing's settings take**, each over the ones before it, field by field (a struct group merges field by field): 1. the schema's declared defaults; 2. the type's own `settingsDefaults`; 3. the first matching Preset Manager row; 4. the creating importer's settings for the type; 5. the values the caller passed. A component takes its `public` defaults, then the row, then the values the caller passed. **Rows change only new things.** A row's values are stamped into what is created while it stands. Editing or removing a row changes nothing that already exists, and opening a world applies nothing to its saved content. A re-import or an `overwrite = true` re-authoring keeps the settings the asset already holds. **The three creation moments, and what each filter reads:** | Moment | The row's declarer | The filter reads | |---|---|---| | `asset.create` | an assetType whose instances share its schema | the new asset's path under `/source` (`sprites/hero.texture`); a runtime asset reads the path the same call would give a durable one | | an import | an importer | the source file's path under `/source` at its first import (`sprites/hero.png`) | | `component.add` | a component | the entity's hierarchy path, names from the root down joined with `/` (`level/props/crate`) | A component takes a row when it is added in edit mode, and in any mode inside a scene build (a scene's `build.luau`, a `SceneModule` placement). A component put back from existing content (a saved scene, a peer's record, a template, a bundle's hierarchy, what a component's own code adds outside a build) takes none. A rebuild restates a component only when its code states it differently, so a row edit alone changes nothing a build already placed. **Checking what applies:** - `preset.defaults()` lists the rows in order, each resolved, with every problem that keeps it from applying. - `preset.defaultFor(declarer, path)` returns the values a new thing at `path` would take from the rows, and the row that gives them. ## In the Inspector Selecting a preset shows, beside its identity: - **Target** — the declarer as an asset chip and the guid the preset names it by, with a notice when the target changed since the preset pinned it or does not resolve. - **Values** — every field of the target's current schema as its typed control, the way the target's own Settings section shows it. An edit is checked against that schema, written with `preset.setValues`, and is one undo step. A stored value the schema refuses is named with the reason. Beneath the fields, **Apply to selection** writes the preset into every selected asset its target governs and into the target component on every selected entity, as one undo step. A preset still in the component-only form shows its values read-only, with a note that its next local edit converts it. Every asset's and component card's Settings section offers **Save as preset**, which captures the values it shows as a preset targeting their declarer and selects it, and **Apply preset…**, which lists the presets targeting that declarer and applies the one picked. ## How the target stays bound - The preset resolves its target by the envelope's guid, and the preset's `.refs` pins that guid together with the target's checksum, so the link survives the target being renamed or moved. The envelope's `name` is never an identity, so the guid is the preset's one link to its target. - Every read re-checks the values against the target's schema as it is now. `report.pinned` is true once `.refs` pins the target; `report.targetChanged` is true while the target's checksum differs from the one pinned, and a field the target dropped or retyped is then rejected by name while the rest still apply. - A `target` written as an identity string is accepted on read, resolves by identity and reads as unpinned; the preset's next local edit rewrites it as the envelope. ## Discovery - `asset.list("preset")` — every registered preset. - `asset.inspect("<name>").detail` — `target`, `targetGuid`, `format`, `values`, `rejected`, `pinned`, `targetChanged`. - `asset.validate("<name>")` — `preset.target_unresolved` when the target does not resolve, `preset.value_rejected` for each value its target rejects. ## Authoring conventions - Name presets by the configuration they capture (`pixel_art`, `hdr_skybox`, `synty`), not by the declarer. - Keep `values` to the fields that define the configuration; fields left out stay as each asset or component has them. ## Common pitfalls - **Hand-edited values.** A value the target's schema rejects is reported, not applied, and every edit of the file warns naming it. - **A table-valued component field** (a `Field.table`) is written whole by `preset.apply`, so a preset holding one provides every key of it. ## Related types - `.assetType` / `.importer` / `.component` — the declarers whose settings a preset holds. - `.bundle` — for multi-component / multi-entity assemblies.

# Shader (asset type) A shader is a WGSL program that runs on the GPU. Materials reference shaders and supply property values; renderable entities reference materials, not shaders directly. This README is the canonical reference for writing **every** kind of shader — it covers all domains, the surface authoring contract, the material property schema, the engine library functions you can call, and how a shader registers and compiles. > New to how shaders, materials, meshes, and textures relate as assets and live > GPU resources? Read `core/resource-model` first. ## Folder shape A `.shader` is a folder asset: ``` <name>.shader/ shader.wgsl # the WGSL body (required) properties.yaml # the material property schema (optional; surface shaders) README.md # a one-paragraph description of THIS shader (required) .metadata # asset metadata ``` Each file carries a committed `.meta` sidecar pinning its stable guid. The identity is `<name>` (the `.shader` suffix strips). A shader's material properties are exactly the ones its `properties.yaml` declares: a shader with no `properties.yaml` declares none, and the engine reads no property from the WGSL body. A `.wgsl` file standing on its own, outside a `.shader` folder, is WGSL source rather than a shader: a material draws with the `.shader` folder that holds it as `shader.wgsl`. Materials and other assets reference a shader by **AssetRef** (a typed, guid-backed handle), never by a raw string. ## Domains The first line of `shader.wgsl` declares the domain: ```wgsl // @domain: surface ``` | Domain | What it is | You write | |---|---|---| | `surface` | A material on a mesh (the common case). | `fragment()` (+ optional `vertex()`/`vertex_clip()`). The engine generates everything else. | | `post-process` | A full-screen pass over the rendered frame, in the space it declares (bloom, vignette, grade). | `fragment(in: PostInput) -> vec4` only. The engine generates the fullscreen `vs_main`/`fs_main` and the group(1) interface from `properties.yaml`. | | `screen` | A standalone full-screen visual (UI background, procedural). | `fragment(in: ScreenInput) -> vec4` only. The engine generates the fullscreen `vs_main`/`fs_main` and the group(1) interface from `properties.yaml`. | | `sky` | The sky background (a fullscreen material). | `fragment(in: SkyInput) -> vec4` only. The engine generates the fullscreen `vs_main`/`fs_main` and the group(1) interface from `properties.yaml`. | | `compute` | GPU data processing, no rasterization. | *(legacy)* An `@compute @workgroup_size(...)` entry over storage buffers — you write the `@group/@binding` lines yourself. **The canonical path is now the separate `.computeShader` assetType** (see below); `.shader` + `// @domain: compute` still compiles for back-compat. | `// @domain:` resolves the names above and `postprocess` / `screenspace` / `ui` as further spellings of `post-process` and `screen`. A body carrying no `@domain:` line is `surface`. `shaderRef:domain()` answers the domain a body declares, a further spelling as the domain it names. Any other name is refused where the shader registers, with an error naming the spellings that resolve, so a shader reaches the renderer only under a domain the compiler scaffolds. Particles are drawn and simulated through these domains like anything else: the particle package's `particle_sprite` and `particle_mesh` are `surface` shaders, and its simulation, culling and sorting steps are `.computeShader` assets. ## Texture coordinate space A shader that samples a texture through a coordinate it builds itself declares which space that coordinate lives in, on a line beside `@domain:`: ```wgsl // @domain: surface // @uv_space: world ``` | Value | The coordinate | A material's `uv_scale` means | |---|---|---| | `mesh` (the default) | The mesh's own UV attribute, `input.uv`. | How many times the texture repeats across a surface. | | `world` | The fragment's world position, as a triplanar projection samples. | How many times the texture repeats per world unit, so a larger surface carries more. | Texture residency reads the declaration to work out how finely a frame samples each texture: a material laying eight copies of its texture across a surface samples each copy at an eighth of the surface's size, and the mip level that serves it is three levels coarser than the surface's own footprint implies. Leaving the line out puts the shader in mesh space, which is what sampling `input.uv` is. `renderer.textureStreaming().textures` reads back the span each texture was measured at and the level that span asked for. **Compute shaders are now their own assetType.** A `.computeShader/` folder gives compute the same zero-scaffolding contract surface/sky/post/screen get: you write only `@compute fn main` (+ helper fns/structs, **no `@group/@binding` lines**) and declare bindings in a sibling `bindings.yaml` (storage buffers, textures, samplers, storage textures, plus a `params:` uniform); the engine generates the `@group(0)/@binding(N)` declarations. Dispatch it through a resolved reference (`asset.resolve("name", "computeShader"):dispatch(...)`). See the `.computeShader` assetType README and `man compute`. The `.shader` compute domain above remains only for back-compat. **Surface**, **sky**, **post-process**, and **screen** shaders all get engine-generated scaffolding — you write only `fragment()` and declare properties in `properties.yaml`; the engine generates the entry points and the group(1) material interface. (Each domain's `fragment` takes a domain-specific input instead of `FragmentData`: `SkyInput`, `PostInput`, or `ScreenInput` — see "Sky shaders" and "Post-process & screen shaders" below.) Compute shaders get the same treatment via the separate `.computeShader` assetType (write only `@compute fn main`, declare bindings in `bindings.yaml`). The legacy `compute` domain here writes its own entry points directly — see `man compute`. ## Surface shaders A surface shader writes **only** the parts that are unique to it. The engine owns all the scaffolding — the vertex stage, every render-mode entry point (`vs_main` / `fs_forward` / `fs_gbuffer`), group(0) (camera, transforms, lights, shadows, and the material rows), and the material interface generated from `properties.yaml` (`material` and your texture slots). Adding a render path never touches your shader. Inside `fragment()` you have access to exactly four things — nothing is hidden: 1. **`input: FragmentData`** — the interpolated per-fragment surface data the engine hands you: world position, world normal, uv, view direction, vertex colour, and more. Every field is in the `FragmentData` table below. 2. **`material.*` and your textures** — the material inputs **you** declare in `properties.yaml`. The mapping is mechanical: a scalar property named `foo` becomes `material.foo`; a texture property named `bar` becomes `bar` (the texture) and `bar_sampler` (its sampler). `material` has **exactly** the fields your `properties.yaml` declares and no others — to add an input, add a property. List what a shader exposes with `shaderRef:getProperties()` or `asset.inspect("<name>")`. 3. **Engine state** — the scene's camera, lights, time, and shadows that the renderer binds for every surface shader. `uniforms.*` (sun direction/colour, ambient, `camera_pos`, `time`) and the punctual lights (`zero_point_light(i)`) are always in scope; the shadow functions come from `#include "@builtin::shaderModules.shadows"`. This is how you light a surface **without** the prebuilt PBR — see "Engine state" and "Writing your own shading" below. 4. **Library functions** you `#include` (see "Library includes"). A shader and its property schema go together. The two files below are a complete surface shader: `properties.yaml` — declares the material inputs: ```yaml properties: - { name: base_color, type: color, default: [1, 1, 1, 1] } - { name: roughness, type: float, default: 0.5, min: 0, max: 1 } - { name: metallic, type: float, default: 0.0, min: 0, max: 1 } - { name: base_color_texture, type: texture, default: white } ``` `shader.wgsl` — those declarations are now `material.base_color`, `material.roughness`, `material.metallic`, and the `base_color_texture` / `base_color_texture_sampler` pair: ```wgsl // @domain: surface #include "@builtin::shaderModules.pbr_shading" // the PBR helpers — see "Library includes" fn fragment(input: FragmentData) -> vec4<f32> { var s = zero_pbr_default_surface(input); s.albedo = material.base_color.rgb * textureSample(base_color_texture, base_color_texture_sampler, input.uv).rgb; s.perceptual_roughness = material.roughness; s.metallic = material.metallic; return vec4<f32>(zero_pbr_shade(input, s), material.base_color.a); } ``` `fragment()` returns the final colour. The engine forces no lighting model on you — flat, toon, PS1, dithered, or PBR are all just different `fragment()` bodies. Calling `zero_pbr_shade` is opt-in. ### The two surface bodies, and what each gives up A surface shader exposes one of two entry points, and which one decides how much of the shader the engine can see: | Body | The engine gets | PBR debug passes read | |---|---|---| | `fn fragment(input: FragmentData) -> vec4<f32>` | the finished pixel, and the `PbrSurface` too when the body hands one to `zero_pbr_shade` | the real channel where a surface was handed over; magenta, the "no data here" marker, where the body shades by its own math | | `fn surface(input: FragmentData) -> PbrSurface` | the surface (albedo, roughness, metallic, emissive, normal, occlusion), which the engine then lights | the real channel | Both are first-class: `fragment()` is what you want when the shader owns its own look, `surface()` when you want engine lighting. `capture { pass = "albedo" }` (and `roughness` / `metallic` / `ao` / `emissive`) answers about your material for either body wherever the engine has the surface — writing the same `PbrSurface` behind `surface()` or behind `fragment()` + `zero_pbr_shade` gives one answer per channel. A body that lights itself by its own math draws those channels magenta. `pbr.shader` exposes `surface()`. `capture { pass = "normal" }` is a geometry read rather than a material one, so it answers for every shader: the normal of the surface where one was handed over, and the interpolated normal of the triangle where none was. It is the same normal the frame's normal buffer carries, which is what a screen-space pass reading `@scene.normal` sees. `shaderRef:shadingModel()` reports which one a shader ended up with — `"engine-lit"` for `surface()`, `"self-shading"` for `fragment()`. ### Optional vertex hooks ```wgsl // Displace geometry in OBJECT space. Runs on skinned meshes too (it sees the // already-skinned, morph-applied local position). Return the modified vertex. // The `v` parameter is immutable (WGSL function params are), so copy it into a // `var` first and modify that. fn vertex(v: VertexData) -> VertexData { var out = v; out.position += out.normal * sin(out.time) * 0.1; return out; } // Post-transform vertex hook (runs AFTER the engine's world/skin/projection). Two // things you can do here, separately or together: // * move `clip_position` — vertex jitter, PS1 snapping, render-resolution lowering; // * write `color` — PER-VERTEX (Gouraud) shading: light the vertex from its // world_normal + the light context and the engine forwards the result into the // interpolated `vertex_color` the fragment reads. fn vertex_clip(v: ClipVertex) -> ClipVertex { var out = v; // Gouraud: Lambert sun + ambient, evaluated once per vertex (flat-banded shading). let n_dot_l = max(dot(v.world_normal, -normalize(v.light_direction)), 0.0); let lit = v.ambient_color * v.ambient_intensity + v.light_color * v.light_intensity * n_dot_l; out.color = vec4<f32>(v.color.rgb * lit, v.color.a); out.clip_position = round(out.clip_position * 80.0) / 80.0; // PS1 snap return out; } ``` #### Declaring how far a hook moves geometry: `vertexBoundsDilation` Static geometry is culled per cluster against its bounds as authored. A hook that moves vertices — wind sway, an outline hull, a vertex wave — can move them outside those bounds, and the cull then drops clusters whose displaced triangles are still on screen. Declare the farthest the hook moves any vertex, in world units, as a `vertexBoundsDilation` float property, with a default that covers the shader's own properties across their ranges: ```yaml properties: - { name: sway, type: float, default: 0.5, min: 0, max: 4 } # 0.4 x sway, at sway's largest value. - { name: vertexBoundsDilation, type: float, default: 1.6, min: 0, max: 100 } ``` Each material sets its own value — a material that sways less can set less, one moved farther sets more — and a material that sets none takes the default. The cull adds the material's value to every bound in its frustum and occlusion tests (not to the level-of-detail cut). A shader with a hook and no `vertexBoundsDilation` property is named once in the engine log when it compiles; its geometry is culled against its undisplaced bounds. ### Structs the engine passes you `FragmentData` (input to `fragment()`): | Field | Type | Meaning | |---|---|---| | `world_position` | `vec3<f32>` | fragment world-space position | | `world_normal` | `vec3<f32>` | interpolated geometric normal (normalized) | | `world_tangent` | `vec4<f32>` | xyz = world tangent, w = handedness (mikktspace) | | `uv` | `vec2<f32>` | primary texture coordinates (perspective-correct) | | `affine_uv` | `vec2<f32>` | the same UVs interpolated **without** perspective correction — the PS1 texture "swim". Always available; sample with it instead of `uv` for affine texture mapping | | `vertex_color` | `vec4<f32>` | interpolated per-vertex colour (white if unset; carries per-vertex lighting if `vertex_clip()` wrote it — see Gouraud below) | | `view_position` | `vec3<f32>` | camera world position | | `view_direction` | `vec3<f32>` | normalized world-space direction to the camera | | `light_direction` | `vec3<f32>` | directional (sun) travel direction, unit length; surface→sun is `-light_direction`. Where no sun reaches the camera it points straight down | | `light_color` | `vec3<f32>` | sun colour | | `light_intensity` | `f32` | sun intensity; 0 where no sun reaches the camera | | `ambient_color` | `vec3<f32>` | ambient colour | | `ambient_intensity` | `f32` | ambient intensity | | `frag_coord` | `vec4<f32>` | framebuffer pixel coords (xy) + depth (z) — for screen-space dither | | `receives_shadows` | `f32` | per-instance: > 0.5 if this surface receives shadows | | `material_index` | `u32` | per-instance: the index standing for the material this surface draws with — the number `renderer.materialIdentity()` maps to a name | | `shader_data` | `array<vec4<f32>, 4>` | per-instance: this entity's own four `vec4` lanes — its `ecs.ShaderData` component's `lane0`..`lane3`, or what `renderer.instanceData.set` put in them (zero until something does). A procedural graph's elements set them through the `data0`..`data3` attributes | | `front_facing` | `bool` | true for front faces (double-sided shading) | | `time` | `f32` | gameplay seconds: advances with the world at `engine.timeScale` and stands still while gameplay is paused (edit mode starts paused), so an effect slows and stops with the world | `shader_data` is the per-instance channel, and it is what lets ONE material serve many entities that differ in a value — a dissolve at its own progress per subject, a hit flash at its own age, a fill level per instance. The alternative is a material per entity, which is a separate pipeline binding and a separate row in the material table for every one of them. ```wgsl const DISSOLVE_LANE: u32 = 0u; fn fragment(input: FragmentData) -> vec4<f32> { let progress = input.shader_data[DISSOLVE_LANE].x; ... } ``` ```lua -- one material on twenty entities, each at its own progress renderer.instanceData.set(subject, 0, progress) renderer.instanceData.clear(subject) -- back to zero lanes ``` The engine attaches no meaning to a lane: a shader and whatever writes it agree what each one carries. Name the indices in the module that writes them. `renderer.instanceData.laneCount()` is how many there are. The same lanes reach a hardware-instanced population through `renderer.mesh.drawInstanced`'s `instanceDataBuffer`, on slots no entity owns. `VertexData` (in/out of `vertex()`): `position` (object-local, model-scale + morph applied), `normal`, `uv`, `color`, `time`. `ClipVertex` (in/out of `vertex_clip()`) — the post-transform vertex. Read context: `clip_position`, `world_position`, `world_normal`, `uv`, `view_position`, `view_direction`, `light_direction`, `light_color`, `light_intensity`, `ambient_color`, `ambient_intensity`, `time`, `view_proj` (the view-projection of the frame being evaluated), and where the vertex stands on its instance: `local_position` (its position in the mesh, after `vertex()` and before the instance's world transform), `object_origin` (the world position of the instance's origin), `object_scale` (the world length of the instance's x, y and z axes) and `object_axes` (`mat3x3<f32>`, the world direction of those axes, unit length) — what a camera-facing card, a shard grown about its own base, or anything placed per instance is built from. **Writable:** `clip_position` (snap / jitter) and `color` (write a lit colour here for per-vertex Gouraud shading; the engine forwards it to the interpolated `vertex_color` — left untouched, the mesh's own vertex colour passes through). `vertex()` and `vertex_clip()` run in the vertex stage, which reads the frame uniforms (`uniforms.*`, and the `ClipVertex` members built from them), the scene buffer (`zero_object_data`, `zero_object_visuals`, the node transforms), the material interface and a `read` buffer property. The point, spot and area lights (`zero_point_light`, `zero_punctual_projection`), the shadow maps, the environment and light probes, the feature textures, the scene depth, colour and reflection copies and a `@group(2)` buffer the shader declares `read_write` are visible to the fragment stage alone: read them in `fragment()` or `surface()`. A shader whose vertex hook reaches one fails to compile, and the error names the stage, the binding and the calls that reach it. ### Each hook runs twice: this frame and the one before it A vertex hook decides where the surface IS, so it also decides where the surface WAS. The engine runs every hook a shader declares a second time, over the same vertex standing one frame back — the previous-frame position stream, with `time` = the clock that frame was drawn at and (for `vertex_clip()`) `view_proj` = the view-projection it was drawn with. The motion vector the renderer publishes is taken between the two results, so a displacement that moves with time reports the velocity of the displaced surface, and a displacement that does not move reports the motion of the transform alone because the identical offset appears on both evaluations and cancels. What that asks of a hook: - **Take the frame from the argument.** `v.time` is the clock of the evaluation being run; `uniforms.time` is always this frame's. A displacement built on `uniforms.time` is identical on both evaluations, so its motion vector cancels to exactly zero. The same holds for `v.view_proj` against `uniforms.view_proj`: a clip hook that rebuilds its position from world space through `uniforms.view_proj` reprojects the previous frame with this frame's camera, and reports no camera motion at all. The wind helpers come in both shapes for this reason — `zero_wind_sway_at_time(world_pos, stiffness, v.time)` inside a hook, `zero_wind_sway(world_pos, stiffness)` outside one. - **Anchor a wrap to `uniforms.time`.** A displacement that jumps discontinuously — a falling drop that restarts at the top, a looping offset — differences its jump as velocity on whichever frame it wraps. Compute the wrap COUNT from `uniforms.time` and the position within the cycle from `v.time`, and both evaluations share the same cycle, so the surface reports the speed it fell at. `@builtin::shaders.rain_material` and `snow_material` do exactly that. The per-vertex attributes (`world_normal`, `uv`, `color`) and the lighting and camera-position context are this frame's on both evaluations, so a hook that displaces along its normal reports the motion of the transform plus this frame's displacement. ## Engine state — point lights & shadows The common lighting state — the sun (`input.light_direction` / `light_color` / `light_intensity`), the ambient (`input.ambient_color` / `ambient_intensity`), the camera (`input.view_position` / `view_direction`), and `input.time` — is already on `FragmentData`, so most shaders never touch the raw engine state. The one thing not on `FragmentData` is the punctual-light array (its length varies), which is always in scope: | Expression | Type | Meaning | |---|---|---| | `uniforms.point_light_count` | `u32` | number of active point/spot lights | | `zero_point_light(i)` | `PointLightData` | one point or spot light (fields below) | | `zero_punctual_projection(i)` | `PunctualProjection` | a spot or area light's projection: `view_proj`, the frustum its cookie and shadow are placed by, and `cookie_layer` | `PointLightData`: `position`, `radius`, `color`, `intensity`, `direction`, `cone_outer_cos`, `cone_inner_cos`, `kind` (0 = point, 1 = spot). Shadows come from `#include "@builtin::shaderModules.shadows"`: - `compute_directional_shadow_factor(world_pos, normal, light_dir) -> f32` — 1.0 lit, 0.0 fully shadowed. - `compute_spot_shadow_factor(...)` / `compute_point_shadow_factor(...)` / `compute_area_shadow_factor(...)` — one shadow shape each. - `compute_punctual_shadow_factor(light_index, world_pos, normal) -> f32` — the row's own shape, read off its `kind`, for a loop that walks the punctual lights and does not want the switch. - `zshadow_light_reach(light_index, world_pos, normal, object_channels) -> f32` — how much direct light that row puts on a surface on those light channels: the channel mask, falloff, cone or rect orientation, and the cosine, times intensity. Zero exactly where the light contributes nothing. ## Writing your own shading The engine forces no lighting model — `zero_pbr_shade` is just one option. You can light a surface yourself from the state above. Here is a complete custom-lit (Lambert diffuse) surface with no PBR: ```wgsl // @domain: surface #include "@builtin::shaderModules.shadows" // exposes compute_directional_shadow_factor fn fragment(input: FragmentData) -> vec4<f32> { let n = normalize(input.world_normal); let albedo = material.base_color.rgb; // Directional (sun) light: Lambert N·L, attenuated by its shadow. let sun = normalize(-input.light_direction); // surface -> sun let sun_shadow = compute_directional_shadow_factor(input.world_position, n, input.light_direction); var lit = input.light_color * input.light_intensity * max(dot(n, sun), 0.0) * sun_shadow; // Point + spot lights (the punctual-light array, varying length). for (var i = 0u; i < uniforms.point_light_count; i = i + 1u) { let pl = zero_point_light(i); let to_light = pl.position - input.world_position; let dist = length(to_light); let l = to_light / max(dist, 1e-4); let atten = max(1.0 - dist / max(pl.radius, 1e-4), 0.0); lit += pl.color * pl.intensity * max(dot(n, l), 0.0) * atten; } // Ambient fill. lit += input.ambient_color * input.ambient_intensity; return vec4<f32>(albedo * lit, material.base_color.a); } ``` - **Unlit:** skip the lights and return your colour directly. - **Toon / PS1 / custom:** start from the loop above and band, quantize, or warp it however you like. - **PBR:** build a `PbrSurface` and call `zero_pbr_shade` (see "Library includes") instead of writing the BRDF by hand. ## Material properties — `properties.yaml` A surface shader declares its material properties in a sibling `properties.yaml`. The engine GENERATES the entire group(1) interface from it — the uniform struct (`material.<name>`) plus a texture/sampler pair per declared texture. You write no `@group`/`@binding` lines. ```yaml properties: - { name: base_color, type: color, default: [1, 1, 1, 1] } - { name: roughness, type: float, default: 0.5, min: 0, max: 1 } - { name: metallic, type: float, default: 0.0, min: 0, max: 1 } - { name: emissive, type: color, default: [0, 0, 0, 1] } - { name: base_color_texture, type: texture, default: white } - { name: normal_texture, type: texture, default: flat_normal } ``` | `type` | WGSL field type | Notes | |---|---|---| | `float` | `f32` | optional `min`/`max` for editor sliders | | `float2`/`float3`/`float4` | `vec2/3/4<f32>` | | | `color` | `vec4<f32>` | sRGB colour, default is `[r,g,b,a]` | | `int` | `i32` | | | `bool` | `u32` | | | `texture` | `texture_2d<f32>` + `sampler` | `default:` is a built-in texture name (`white`, `black`, `flat_normal`) | In the shader you read scalars as `material.<name>` and sample textures as `textureSample(<name>, <name>_sampler, uv)`. A texture slot with no bound texture samples its default, so a shader can always sample — no presence flag needed. `vertex()` reads the same material interface, so a height or deformation map drives displacement from the slot it was bound to. A vertex stage has no implicit derivatives, so sample it at an explicit mip there: `textureSampleLevel(<name>, <name>_sampler, uv, 0.0)`. ## Sky shaders A `// @domain: sky` shader draws the sky as a fullscreen pass. You write **only** `fn fragment(in: SkyInput) -> vec4<f32>` — the colour along the view ray for each pixel — and declare tunables in `properties.yaml` exactly like a surface shader (read as `material.<name>`, textures via `textureSample(<name>, <name>_sampler, uv)`). The engine generates the fullscreen `vs_main`/`fs_main`, the group(1) interface, and reconstructs the world-space ray; there is no separate sky renderer — a sky material runs through the same material pipeline, just on a fullscreen triangle. A sky pipeline binds group(0) and the group(1) interface alone, so a sky that reads a `buffer` property or any other `@group(2)` binding fails to compile, and the error names the binding and the calls that reach it. `SkyInput` (the only argument) carries: | Field | Type | Meaning | |---|---|---| | `ray_dir` | `vec3<f32>` | normalized world-space ray through this pixel | | `ndc` | `vec2<f32>` | screen position in NDC `[-1, 1]` | | `sun_direction` | `vec3<f32>` | unit vector TOWARD the sun (from the directional light) | | `camera_pos` | `vec3<f32>` | world-space camera position | | `time` | `f32` | gameplay seconds: advances with the world at `engine.timeScale` and stands still while gameplay is paused (edit mode starts paused), so an effect slows and stops with the world | ```wgsl // @domain: sky fn fragment(in: SkyInput) -> vec4<f32> { let up = max(in.ray_dir.y, 0.0); let col = mix(material.horizon_color.rgb, material.zenith_color.rgb, up); return vec4<f32>(col * material.exposure, 1.0); } ``` Day/night comes from `sun_direction.y` (the directional light's elevation), so a sky shader never needs a time-of-day uniform — orient the light and the sky follows. The builtin sky shaders (`procedural_sky`, `solid_sky`, `cubemap_sky`, `equirect_sky`) are all authored this way; author a custom sky material and point a `Skybox` component at it. ## Post-process & screen shaders `// @domain: post-process` and `// @domain: screen` shaders are authored exactly like surface and sky — you write **only** `fn fragment(in: PostInput) -> vec4<f32>` (post) or `fn fragment(in: ScreenInput) -> vec4<f32>` (screen) and declare tunables in `properties.yaml` (read as `material.<name>`, textures via `textureSample(<name>, <name>_sampler, uv)`). The engine generates the fullscreen `vs_main`/`fs_main` and the group(1) material interface. **There are no `uniforms.params[i]` slots** — that convention is retired; every input is a named property. A **post-process** shader's `properties.yaml` carries one further, top-level key beside `properties:` — `space`, either `light` (the default) or `display`: ```yaml space: display properties: - { name: amount, type: float, default: 1.0, min: 0, max: 1 } ``` `light` is the scene's linear radiance — what every surface, the sky and every `afterLighting` feature write, unbounded above 1.0. `display` is the code values the `__display__` pass leaves behind after applying exposure and its tone curve — the 0-1 range a colourist's tool and a dithered or quantized effect expect. `postprocess.add`'s `opts.space` overrides what the asset declares for one registration. The chain converts at both ends of an effect that runs on a segment in the other space, so a light-space vignette still darkens a frame graded after `__display__`, and a lookup table built from a `.cube` export always sees the code values it was authored against, wherever the chain runs it. `space` also decides which layer a registration that names none lands on: a `light` shader grades `scene`, the segment holding the scene's own light, and a `display` shader grades `all`, the picture after `__display__` with the UI on it. So `postprocess.add("glow", asset.resolve("~.shaders.bloom", "shader"))` thresholds highlights that still stand above 1.0, and a lookup table registered the same way reads the code values it was built for. `opts.layer` names one explicitly — `"all"` is what an effect that must grade the interface says. A **post-process** shader's `properties.yaml` carries a further, optional top-level key — `passes:`, the pass list the frame draws through, in order: ```yaml space: light passes: - { name: bright, entry: bright, scale: 0.5, inputs: [scene] } - { name: down1, entry: downsample, scale: 0.25, inputs: [bright] } - { name: up0, entry: upsample, scale: 0.5, inputs: [down1, bright] } - { name: composite, entry: composite, inputs: [scene, up0] } ``` Every pass's entry — `fn <entry>(in: PostInput) -> vec4<f32>` — lives in the one `shader.wgsl` the assetType recompiles on. `scale` is the fraction of the view a pass draws at; the LAST pass takes none, since it draws at the view's size and writes the chain's output. `inputs` names what a pass reads — `"scene"`, the chain's input for this effect, or the name of an EARLIER pass — and `inputs[0]` is what `zero_sample_scene(uv)` reads inside that pass's body; every input, `inputs[0]` included, is also reachable by index through `zero_pass_input(i, uv)` and `zero_pass_input_size(i)`, below. The one place the two reads of `inputs[0]` disagree: in an effect whose declared space differs from the segment it draws on, `zero_sample_scene` converts a `scene` input into that space while `zero_pass_input(0, uv)` hands back the segment's own, unconverted value. A shader declaring no `passes:` is one pass whose entry is `fragment`, and its generated WGSL is byte-identical to what it is without the key. A pass that is not the last writes a scratch target the effect owns — in the space the shader declared, view-sized times its own `scale` — allocated on the first draw and replaced whenever the size changes. `engine.params`, below, carries the drawing pass's own index, its scale, and its first input's texel size, so one shader body serves every level of a chain built this way. The worked example is the bloom: `@builtin::shaders.post.bloom` declares twelve passes — a threshold pass, five downsamples, five tent upsamples and a composite — out of four entries, since the five downsamples share one and the five upsamples share another, each level reading the one below it through `zero_pass_input` and telling itself apart by `engine.params`. Its own README walks the declaration. A declaration is refused at compile time — reported through `compileStatus()`, `postprocess.add`'s second return value, or a live effect's `status()[i].error` — for: an entry `passes:` names that `shader.wgsl` does not define; an input naming a later pass, or naming neither `scene` nor an earlier pass; a `scale` outside `(0, 1]`; a `scale` declared on the last pass; a `buffer` property (group(2) carries the pass inputs now); or a shader whose group(0) plus declared texture slots plus widest pass's inputs ask for more than the 16 FRAGMENT sampled textures one shader stage has. A **post-process** shader runs as a full-screen pass over the already-rendered frame. Its `fragment` argument and framework helpers: `PostInput`: | Field | Type | Meaning | |---|---|---| | `frag_coord` | `vec4<f32>` | framebuffer pixel coords (`@builtin(position)`) | | `uv` | `vec2<f32>` | screen UV, `(0,0)` top-left … `(1,1)` bottom-right | Framework helpers (always in scope — no include needed): | Call | Returns | Meaning | |---|---|---| | `zero_sample_scene(uv)` | `vec4<f32>` | the rendered scene colour at `uv`, converted into the space this effect declared (`zero_scene_light` is the same call, for content that reads it by that name) | | `zero_scene_depth(uv: vec2<f32>)` | `f32` | scene depth at `uv` — a screen UV here, where the surface domain's `zero::scene_depth` module takes a pixel coordinate under the same name (exact texel — an R32Float copy). Reversed: 1.0 is the near plane, 0.0 the far plane and empty sky, so a nearer surface reads GREATER and `d > 0.0` means "a surface was drawn here" | | `zero_linear_depth(d, near, far)` | `f32` | that sample as a distance in metres — the one inversion of the reversed range | | `zero_scene_motion(uv)` | `vec2<f32>` | per-pixel screen-space motion vector at `uv` | | `zero_pass_input(i, uv)` | `vec4<f32>` | the pass's declared input `i`, at `uv`. Input 0 is what `zero_sample_scene` reads | | `zero_pass_input_size(i)` | `vec2<f32>` | that input's size in texels, so one entry serves every level of a chain | Engine data uniform `engine` (a `PostEngineData`): | Field | Type | Meaning | |---|---|---| | `resolution` | `vec4<f32>` | `x`=width, `y`=height, `z`=1/width, `w`=1/height | | `time` | `vec4<f32>` | `x`=elapsed seconds, `y`=dt, `z`=frame index | | `view_proj` | `mat4x4<f32>` | this frame's view-projection | | `prev_view_proj` | `mat4x4<f32>` | last frame's view-projection (motion) | | `inv_view_proj` | `mat4x4<f32>` | inverse view-projection (world-space reconstruction) | | `params` | `vec4<f32>` | `x`=this pass's own index, `y`=the scale it draws at, `zw`=1/size of its first input — for a pass the post-process chain draws. A pass a `renderFeature` fragment pass draws carries that feature's `spec.params` here instead; a shader declaring `passes:` is always drawn by the chain, so the two never meet in one shader. | The engine fills all of it **per render target**. A registered effect runs over the live viewport's frame AND over every offscreen one — a `capture` from a world-space station or orbiting an entity, a capture of a named camera, a render-to-texture camera — and in each of those, `view_proj`, `prev_view_proj` and `inv_view_proj` are the camera THAT render was drawn from, while `resolution` is that target's own pixel size. So a pass reconstructing world space from `zero_scene_depth(uv)` reconstructs against the station and lens the capture asked for, and an offscreen frame is a picture of the effect as it stands at that station. (A render feature's passes reach an offscreen render the same way, against the same camera — the `renderFeature` assetType README carries that half, including `@frame.camera`, its own per-target camera input.) An offscreen render keeps no view history of its own, so `prev_view_proj` there holds that same matrix rather than the frame before it, and a pass taking camera motion from the two reads none. On the viewport the pair is a frame apart, so camera motion is the one reading a capture answers differently from the screen. `postProcessing = false` on a capture is the one control that takes the chain off the frame it returns; it names the camera property the ephemeral capture camera is built with, so it scopes to that one capture. What a capture opts out of is the registered effects; `__display__` still runs on it, so the exposure and tone curve that turn the scene's light into a picture are on what a capture returns. A diagnostic capture (`albedo`, `normal`, `motion_vectors`, …) bypasses `__display__` itself, since a channel like that is a reading of the scene rather than a graded picture of it. The environment-cube capture bypasses it too, and for the same reason: a probe face carries radiance into the cube array, and `__display__` would fit that radiance to a display's code values — a picture of it rather than the value the PBR indirect term multiplies into its lighting. A complete post-process effect — a vignette via `fragment()` + `properties.yaml`: `properties.yaml`: ```yaml properties: - { name: intensity, type: float, default: 0.5, min: 0, max: 1 } - { name: radius, type: float, default: 0.75, min: 0, max: 1 } - { name: softness, type: float, default: 0.45, min: 0, max: 1 } ``` `shader.wgsl`: ```wgsl // @domain: post-process fn fragment(in: PostInput) -> vec4<f32> { let color = zero_sample_scene(in.uv); let intensity = max(material.intensity, 0.01); let radius = max(material.radius, 0.01); let softness = max(material.softness, 0.01); let center = in.uv - vec2<f32>(0.5); let dist = length(center); let vignette = smoothstep(radius, radius - softness, dist); let darkened = color.rgb * mix(vec3<f32>(1.0), vec3<f32>(vignette), intensity); return vec4<f32>(darkened, color.a); } ``` The asset above **defines** the effect. It enters the frame when something registers it: ```luau postprocess.add("vignette", asset.resolve("~.shaders.vignette", "shader"), { priority = 200 }) ``` The registration names no layer, so the effect grades the layer its `space` belongs to: this vignette reads light, so it runs on `scene`, over the picture before the UI reaches it. Registering it is the one step that makes it render, so a world's chain holds exactly the passes its content asked for. Editing the shader body afterwards recompiles every effect registered from it, in place — the pass keeps its enabled state, priority, layer, tuned property values and bound textures, and draws the new pixels on the next frame. `postprocess.remove(name)` takes it back out of the chain and leaves the asset alone. Drive the named properties at runtime with `postprocess.setProperty(name, prop, value)` and bind texture properties with `postprocess.setTexture(name, prop, path)` (both keyed by the **property name**, never a slot index — see the post-process section of `man shaders` / the `postprocess.*` API). Read them back with `postprocess.describe(name)`: it carries the chain state `postprocess.status()` lists for that effect, the `properties` schema it was registered with, and the `values` each of those properties currently holds. Before the effect exists, `shaderRef:getProperties()` on the asset names the same schema, so what a pass takes is answerable either side of the registration. A **screen** shader is a standalone full-screen visual (UI background, procedural pattern). `ScreenInput` mirrors `PostInput` (`frag_coord` + `uv`); there is no scene to sample, so it has no `zero_sample_scene` helpers. Its `engine` uniform is a `ScreenEngineData`: | Field | Type | Meaning | |---|---|---| | `time` | `f32` | gameplay seconds: advances with the world at `engine.timeScale` and stands still while gameplay is paused (edit mode starts paused), so an effect slows and stops with the world | | `scroll` | `f32` | UI scroll offset | | `resolution` | `vec2<f32>` | render-target size in pixels | | `mouse` | `vec2<f32>` | mouse position | | `velocity` | `f32` | scroll velocity | ```wgsl // @domain: screen fn fragment(in: ScreenInput) -> vec4<f32> { let t = engine.time * 0.5; let col = vec3<f32>( sin(in.uv.x * 3.14159 + t) * 0.5 + 0.5, sin(in.uv.y * 3.14159 + t * 0.7) * 0.5 + 0.5, material.brightness ); return vec4<f32>(col, 1.0); } ``` ## Library includes A library module is a **`.shaderModule` asset**: it only EXPOSES functions and structs and does nothing on its own (exactly like a Unity `.cginc`). You must call what you include. Include one by its identity, the same string you would hand `asset.resolve`: | Include | Exposes | |---|---| | `@builtin::shaderModules.pbr_shading` | `struct PbrSurface { albedo, normal (world), metallic, perceptual_roughness, emissive, occlusion }`; `zero_pbr_default_surface(FragmentData) -> PbrSurface`; `zero_pbr_shade(FragmentData, PbrSurface) -> vec3<f32>` (the full forward PBR lighting) | | `@builtin::shaderModules.shadows` | `compute_directional_shadow_factor(...)`, `compute_spot_shadow_factor(...)`, `compute_point_shadow_factor(...)`, `compute_area_shadow_factor(...)`, `compute_punctual_shadow_factor(...)`, `zshadow_light_reach(...)` | | `@builtin::shaderModules.engine_bindings` | the group(0) declarations (camera, transforms, lights, shadow maps) — surface shaders get this implicitly; include it directly only for advanced raw access | | `@builtin::shaderModules.shadow_bias` | `zshadow_offset_receiver(...)`, `zshadow_directional_depth_bias(...)` and the rest of the bias math, as pure functions | | `@builtin::shaderModules.texturing` | `ztex_stochastic(...)` (tile without a visible repeat grid), `ztex_blend_normals(...)` (reoriented normal blending) | | `@builtin::shaderModules.depth` | `zero_linear_depth(depth, near, far)` — the one reading of a stored depth | | `@builtin::shaderModules.gbuffer_material` (`zero::gbuffer_material`) | `struct ZeroGbufferMaterial { base_color, metallic, perceptual_roughness }`; `zero_gbuffer_unpack_material(texel: vec4<u32>) -> ZeroGbufferMaterial`; `zero_gbuffer_material(t: texture_2d<u32>, coord: vec2<i32>) -> ZeroGbufferMaterial`, `zero_gbuffer_base_color`/`zero_gbuffer_metallic`/`zero_gbuffer_roughness(t, coord)` — the one unpacking of `@scene.material`'s `Rgba8Uint` texel into what a surface reflects | | `zero::scene_depth` | `zero_scene_depth(px: vec2<f32>)`, `zero_scene_depth_uv(uv: vec2<f32>)`, `zero_scene_depth_texel(px: vec2<i32>)`, `zero_scene_depth_size()`, `zero_scene_depth_is_sky(d: f32)`, `zero_scene_world_position(px: vec2<f32>)`, `zero_scene_world_normal(px: vec2<f32>)`, `zero_scene_depth_fade(px: vec2<f32>, world_position: vec3<f32>, fade_distance: f32)` — every `px` is a pixel coordinate in two components, so a surface fragment passes `input.frag_coord.xy`, not the `vec4<f32>` `FragmentData.frag_coord` holds. The depth already drawn UNDER a surface fragment, and the world position and geometric normal of the surface standing there. What a decal projector box, a soft particle, a forcefield or water intersection foam shades against. Surface domain; the depth belongs to the view being drawn | | `zero::scene_color` | `zero_scene_color(px: vec2<f32>)`, `zero_scene_color_uv(uv: vec2<f32>)`, `zero_scene_color_texel(px: vec2<i32>)`, `zero_scene_color_size()`, `zero_scene_color_refract(px: vec2<f32>, offset_px: vec2<f32>)`, `zero_scene_color_refract_uv(px: vec2<f32>, offset_fraction: vec2<f32>)` — every `px` is a pixel coordinate in two components, so a surface fragment passes `input.frag_coord.xy`, not the `vec4<f32>` `FragmentData.frag_coord` holds. The scene colour already composed BEHIND a surface fragment, and the displaced read that bends it. What heat haze, a refraction shockwave, a cloaking or shield warp, glass and ice shade with. Surface domain; the colour belongs to the view being drawn, and is the lit opaque scene as it stood immediately before the transparent pass — so a `blend`-queue surface reads this frame's, an opaque one reads the previous frame's, and two overlapping blended surfaces both read the same picture and do not bend each other | `asset.list("shaderModule")` enumerates the `.shaderModule` assets in this engine, builtin and world-authored alike, and each one's README says what it exposes. Those are the `@builtin::shaderModules.*` rows above; a module you write is included exactly the same way, by its own identity or by the package-relative `~.name` form. The `zero::`-named rows come from the renderer itself. Their WGSL is baked into the engine binary and registered under those names before any asset exists, so they answer an `#include` in every engine and appear in no asset index. A shader that includes one records no dependency on it: the engine reading the shader is the engine that already holds the module. Two spellings reach the same `.shaderModule`: an earlier name is kept working through the module's declared alias, so a shader written against `zero::pbr` still compiles and still pins the asset that name resolves to. Newly written shaders should use the identity — it says which asset travels with the content. Whichever punctuation an include is written with — `#include "name"`, `#include 'name'` or `#include <name>` — names the same module and records the same dependency. The engine's own scaffolding — `zero::surface`, `zero::sky`, `zero::post`, `zero::screen` — is injected by the compiler around the matching domain and is never written by hand. The surface framework puts `safe_normalize`, `mat3_from_mat4` and `apply_normal_map(FragmentData, tangent_normal)` in scope for every surface shader. ## Shader features — compile-time variants A **property** is runtime uniform data (a colour, a float, a texture): set per material, read as `material.<name>`, no recompile. A **feature** decides which code is in the program at all. Use a property when the value changes at runtime, a feature when it changes what the shader is. Declare features in `properties.yaml`, alongside the properties: ```yaml properties: - { name: base_color, type: color, default: [1, 1, 1, 1] } features: - { name: PARALLAX, default: false, description: silhouette parallax march } - { name: DETAIL, default: true, description: detail-normal overlay } ``` Guard the code each feature owns with `#ifdef`: ```wgsl fn surface(input: FragmentData) -> PbrSurface { var s = zero_pbr_default_surface(input); #ifdef PARALLAX s.normal = parallax_normal(input); // absent when PARALLAX is off #else s.normal = input.world_normal; #endif return s; } ``` `#ifdef` / `#ifndef` / `#else` / `#endif`, one per line, and they nest. The directives resolve before the WGSL reaches the compiler, so the program holds exactly the lines its feature set selected and every other line is an empty one, which keeps a compiler message on the line number the file shows. An unbalanced block is a compile error naming its line. A material is what chooses a feature set, so `features:` shapes the shaders a material draws with — the surface, sky and screen-space domains. A material names the features it wants in its `mat.yaml`: ```yaml name: worn_rock shader: "@builtin::shaders.pbr_layered" features: PARALLAX: true # on, over the shader's default DETAIL: false # off, against the shader's default ``` or through `asset.create`: ```lua asset.create("material", "worn_rock", { shader = "@builtin::shaders.pbr_layered", features = { PARALLAX = true, DETAIL = false }, }) ``` ### A property that switches a feature on A feature can also be switched on by a material's property values. List the properties under the feature's `properties:`, and the feature is on for every material whose value for one of them stands away from that property's declared default: ```yaml properties: - { name: clearcoat, type: range, default: 0.0, min: 0, max: 1 } features: - { name: CLEARCOAT, properties: [clearcoat], description: the coat lobe } ``` ```wgsl #ifdef CLEARCOAT s.clearcoat = material.clearcoat; #endif ``` A material at `clearcoat: 0` draws a program holding none of the coat's code; one that sets `clearcoat: 0.6`, in its `mat.yaml`, through `asset.create` or with `renderer.material.setProperty`, draws the `CLEARCOAT` variant, and setting it back to 0 returns it to the base program. The property stays a runtime value the program reads; the feature decides only whether the code reading it is there. This is how a shader with costly optional parts keeps every material that leaves them at rest on the cheap program without each material having to name a feature. A material naming the feature off under `features:` keeps it off whatever the property holds. Write the guarded code so that, with the feature off, the shader computes what it would with the property at its default: the material sees one look either way, and pays for the part only where it shows. Each distinct set of switched-on features is a program to compile the first time a material reaches it. A material keeps the programs its values switched it onto compiled while it lives, so a property crossing back and forth reuses them; when the shader's variant budget is spent, the ones no material draws with are given back first. A `properties:` entry is a value property (`float`, `range`, `int`, `bool`, `color`, `float2` to `float4`) the shader declares; one naming no declared property, or a texture or buffer, is refused by name. The engine compiles one program per feature set its materials ask for, named `<shader>#<features>`. Materials sharing a set share that program. A feature the shader does not declare selects nothing, so a stray name cannot make the engine compile anything, and the compile names both halves of a mismatch: a guard on a name `features:` does not declare, and a declared name no guard tests. **Permutations are the cost.** A shader with n features can express 2^n programs, each its own compile and pipeline. One shader holds at most 32 variants; a material asking for a set past that draws with the shader's default features and the engine logs which shader ran out. Read what a scene is spending: ```lua for _, s in renderer.shaderVariants() do print(s.shader, #s.variants, "of", s.budget) print(renderer.variantSource(s.base.program)) -- what it ships as for _, v in s.variants do print(v.program, renderer.variantSource(v.program)) -- what one asked for end end ``` A shader registers under every key it answers to — its identity, its guid, each alias — and each key is a row. A variant lands on the row for the key the material named, which for authored content is the guid, so take a program name from the report rather than composing one. ## Materials reference shaders A `.material` applies a shader to entities. Its `mat.yaml` references the shader by its identity string and supplies property values: ```yaml name: my_material shader: "@builtin::shaders.pbr" render: blend: alphaBlend cull: back depth_write: false floats: roughness: 0.5 colors: base_color: { r: 1.0, g: 1.0, b: 1.0, a: 1.0 } textures: base_color_texture: { ref: "default:white" } ``` `shader:` is a plain identity string, and each `colors:` entry is an `{ r, g, b, a }` map — the two forms `mat.yaml` accepts for those keys. A key's value is the rest of its line, so every value above stands alone on one. ### The render block Render behaviour that used to be material "flags" — the blend equation, alpha cutout, which faces are drawn — lives in the `render:` block and drives the pipeline, not the fragment. These are the keys that reach the pipeline from both authoring routes, and the values each one takes: | Key | Values | Sets | |-----|--------|------| | `blend` | `opaque` `transparent` `alpha` `alphaBlend` `additive` `add` `cutout` `alphaCutout` `premultiplied` `premultipliedAlpha` `multiply` | the equation the fragment composites with | | `cull` | `back` `front` `none` `off` `disabled` | which faces the rasterizer keeps | | `depth_write` | `true` `false` | whether the draw writes depth | | `depth_compare` | `less` `lessEqual` `lequal` `equal` `greater` `greaterEqual` `gequal` `always` `never` | the test a fragment's depth passes | | `topology` | `triangleList` `lineList` `lines` `line` | what consecutive vertices assemble into | | `queue` | a whole number: `2000` opaque, `3000` transparent, `4000` overlay | the draw order of the material; a number inside a queue orders within it | `blend` is also spelled `type`, and the two name one key. Its equations, over the fragment the shader returned (`src`) and the value already in the target (`dst`): `opaque` is `src`; `transparent`, `alpha` and `alphaBlend` are `src * src.a + dst * (1 - src.a)`; `additive` and `add` are `src + dst`, the mode a glow, a light shaft or a spark card is written for; `premultiplied` and `premultipliedAlpha` are `src + dst * (1 - src.a)`, taking a source that already carries its own coverage; `multiply` is `src * dst`, which is how a soot, grime or shadow card darkens what stands behind it. `cutout` and `alphaCutout` keep the fragment or discard it on its alpha and go on writing depth. The same block is what `asset.create`'s `render` argument takes, key for key and value for value, so one material named in Luau and the same material standing on disk ask for their pipeline with the same words. `depth_write` is the one key whose value takes the shape of its route: the word `false` on a `mat.yaml` line, the boolean `false` in the Luau table. ```lua asset.create("material", "heat_plume", { shader = "@builtin::shaders.pbr", render = { blend = "alphaBlend", cull = "none", depth_write = false }, }) ``` Where a value outside a key's set is reported is what the two routes differ on. `asset.create` refuses the call on any of these keys and names the set it measured the value against, or for `queue`, that it takes a number. A `mat.yaml` standing on disk registers either way, each key falling back to a value of its own: `blend` to `opaque`, `cull` to `back`, `depth_write` to `true`, `depth_compare` to `less`, `topology` to `triangleList` and `queue` to `2000`. The engine log carries the blend fallback, naming the value and the set it was measured against, so a `mat.yaml` value is worth putting through `asset.create` once — that is where every key answers with its set. ## Registration & hot reload A `.shader` is compiled by the asset system, not by the renderer: writing `shader.wgsl` or `properties.yaml` fires the `.shader` assetType's `onChange`, which calls `shader.compile` — that reads the body + the property schema, generates the group(1) interface and (for surface) every render-mode entry point, expands `#include`s, **validates the result with naga**, and registers it. Validation happens at this compile step — the whole point of "compile". A shader with invalid WGSL (an unknown field, a type mismatch, an undefined function) **fails to compile** with the real reason — readable directly from Luau via `shaderRef:compileStatus()` (`{ status = "compiled" | "failed" | "pending", error?, pending? }`), and also in the author log and `asset.inspect`; it is **not** registered. This fires on create and on every edit, so saving a `.wgsl` recompiles with no world reload; the change is visible on the next frame. Compilation is async, so a `compileStatus()` of `"pending"` immediately after a write means check again next frame. A surface shader is also checked for the material resolve, the pass that shades each pixel of the visibility buffer through the shader's own code. That check reaches its verdict a few frames after the raster compile. Until it does, `compileStatus()` reads `"pending"` with `pending` saying so, and the shader already draws. A shader the check refuses reads `"failed"`, and `error` names the construct the resolve cannot evaluate and where it sits; its pixels resolve through the error shader until it changes. The stage the compile checks comes from one of two places: the hook the engine scaffolds entry points around (`fragment()`, or `surface()` for an engine-lit surface), or the `vs_main` / `fs_main` pair the body writes itself. The pair is a pair: a body that writes its own entry points owes both stages. A body carrying neither route — a hook whose name is a letter off, say, or half a pair — holds no stage the rasterizer can enter, so the compile **refuses it** and names both the entry point it looked for and what the body carries, in the same `compileStatus()` error every other compile failure reads back through. ### Reading a compile error against the text it is about What the compiler reads is not the file you wrote. The engine generates the group(1) material interface from `properties.yaml`, wraps the body in its domain's framework and every entry point that domain needs, expands every `#include`, and resolves every `#ifdef` — so a 100-line body reaches naga as tens of thousands of lines. **Every line number in a compile error, and every `[N]` handle index naga prints where it has no name to print, is a position in that composed module.** A fault at module scope (`no definition in scope for identifier: 'zero_scene_depth'`) reads fine without it; a fault inside a function body does not. `renderer.compiledSource(name)` returns that composed module, and `shaderRef:compiledWgsl()` returns it for a ref you already hold. Both answer for a shader that declares no features, needs no material, entity or draw, and — the case they exist for — **answer for a shader whose compile failed**, holding the text the compiler was refused over for as long as `compileStatus()` reports that failure. `renderer.compiledShaders()` lists every name they answer for. ```lua local ref = asset.resolve("myShader", "shader") local status = ref:compileStatus() if status.status == "failed" then print(status.error) -- the line, the caret, and the cause local wgsl = ref:compiledWgsl() -- the text those positions are in local line = tonumber(string.match(status.error, "wgsl:(%d+):")) print(string.split(wgsl, "\n")[line]) end ``` The composed text carries every `zero::*` module the shader included, so the signature a call was rejected against is in the same text as the rejection — which is the whole of what a "wrong argument type" error needs to be actionable. A material bound to a broken (or missing) shader renders with the **magenta error shader** — never invisibly — so a shader problem is impossible to miss on screen as well as in the log. **Fixing the shader and re-saving recovers it live**, with no engine restart: the recompile clears the failed-shader state and the object returns to normal on the next frame. ## Discovery - `asset.list("shader")` — every registered shader. - `asset.inspect("<name>")` — domain, status, source, parse errors, this README. - `shaderRef:compileStatus()` — did it compile? `{ status, error?, pending? }`, no log needed. `"pending"` with `pending` set means the material resolve check is still reaching its verdict; a resolve refusal reads as `"failed"`. - `shaderRef:compiledWgsl()` / `renderer.compiledSource("<name>")` — the WGSL the compiler received, which is what a compile error's positions index into. Answers for a failed compile too. - `renderer.compiledShaders()` — every name that call answers for. - `shaderRef:getProperties()` — the parsed `properties.yaml` schema. - `shaderRef:listMaterialsUsing()` — materials bound to this shader. ## Common pitfalls - **A compile error shows up in one of three ways at the draw.** A shader that never compiled has no pipeline, so the renderer binds the magenta placeholder. A shader that compiled once and was then edited into brokenness KEEPS the pipeline that build produced — the surface goes on drawing the last good program, pixel for pixel, and looks entirely intact while the source is broken. A geometry-pass-only shader has no magenta tell either, since it is not on a forward mesh. `shaderRef:compileStatus()` gives the status + error text (as does `asset.inspect("<name>")`), and `renderer.drawDiagnostics()` names which renderables are in which of the three states — a row carries `substituted` for the placeholder, `stale` for a program that no longer compiles, and the cause either way. - **A compile error's line numbers are lines of the composed module, not of your file.** `renderer.compiledSource("<name>")` / `shaderRef:compiledWgsl()` is that module; resolve the position against it rather than counting lines in `shader.wgsl`. The `[N]` a message prints in place of a name is a naga handle index — it moves with the size of the body and is the same value for two unrelated faults in one file, so it is neither a line number nor a way to tell two failures apart. - **Don't declare group(0) or group(1) in a surface shader.** The engine owns group(0) and generates group(1) from `properties.yaml`. Re-declaring them is the fastest way to a layout mismatch. - **`#include` does nothing by itself** — it only exposes functions; you must call them. - **Renaming a property** changes the generated uniform field; materials setting the old name silently fall back to its default. - **Compute shaders don't render** — verify output via a buffer readback or a debug pass. Author them as a `.computeShader` (its own assetType), not a `.shader` — see `man compute`. ## Related types - `.material` — the required wrapper that applies a shader to entities. - `.style` — UI cascade, not GPU shaders. ## In the Inspector Selecting a shader opens its **Shader** section: its domain and shading model (and, for a post-process shader, the space its `fragment()` reads and returns), whether it compiled (and the compiler's error when it did not), a table of its declared properties (type, default, range), its post-process passes when it has any, the bindings it declares, and openers for its source. The **Materials** section (collapsed at first) has a Find materials action that lists, as links, every material drawing with the shader; it scans every material, so it runs only when asked.