Interface: UseVisionCameraWebGpuSourceOptions

Defined in: react-native-vision-camera-source/src/webgpu/useVisionCameraWebGpuSource.ts:37

Options for useVisionCameraWebGpuSource. Also accepts every VisionCamera frame-output option except pixelFormat, which the hook forces to 'native' (the zero-copy camera-import path requires it).

Extends

• Partial<Omit<FrameOutputOptions, "pixelFormat">>

Properties

cameraShaderBindings?

optional cameraShaderBindings: CameraShaderBindings

Defined in: react-native-vision-camera-source/src/webgpu/useVisionCameraWebGpuSource.ts:62

Camera shader bindings built with createCameraShaderBindings. When set, the render context carries a ready-made cameraBindGroup for the live camera texture every frame.

---

device?

optional device: GPUDevice

Defined in: react-native-vision-camera-source/src/webgpu/useVisionCameraWebGpuSource.ts:57

Bring your own GPUDevice instead of the shared one from useCameraWebGpuDevice. It is validated against the required camera-import features; a device missing any of them surfaces a descriptive error instead of failing per frame.

---

enabled?

optional enabled: boolean

Defined in: react-native-vision-camera-source/src/webgpu/useVisionCameraWebGpuSource.ts:42

Whether the source is live. While false, no track or surface pool exists and nothing is published — the declarative sibling of VisionCamera's isActive. Defaults to true.

---

frameIntervalNanoseconds?

optional frameIntervalNanoseconds: number

Defined in: react-native-vision-camera-source/src/webgpu/useVisionCameraWebGpuSource.ts:77

Fallback spacing between published frames, in nanoseconds, used only when a camera frame carries no usable timestamp. Defaults to 33,333,333 (30 fps).

---

height

height: number

Defined in: react-native-vision-camera-source/src/webgpu/useVisionCameraWebGpuSource.ts:46

Height of the published video, in pixels.

---

onFrame()

onFrame: (frame, render) => void

Defined in: react-native-vision-camera-source/src/webgpu/useVisionCameraWebGpuSource.ts:70

Worklet called for every camera frame. Call render(...) at most once to draw this frame's output; skipping it drops the frame (nothing is published for it). After render(...) returns you may keep using frame (for example run inference) — but only until this callback returns, when the hook releases the frame. Do not retain it. Keep the function's identity stable (useCallback or module scope).

Parameters

Parameter	Type
frame	Frame
render	WebGpuFrameRenderFunction

Returns

void

---

onFrameDropped()?

optional onFrameDropped: (reason) => void

Defined in: react-native-vision-camera-source/src/webgpu/useVisionCameraWebGpuSource.ts:72

Called whenever the camera pipeline drops a frame; forwarded to VisionCamera.

Parameters

Parameter	Type
reason	FrameDroppedReason

Returns

void

---

poolSize?

optional poolSize: number

Defined in: react-native-vision-camera-source/src/webgpu/useVisionCameraWebGpuSource.ts:51

Number of in-flight output surfaces (a pushed frame may still be encoding while the next one is drawn). Defaults to 3.

---

width

width: number

Defined in: react-native-vision-camera-source/src/webgpu/useVisionCameraWebGpuSource.ts:44

Width of the published video, in pixels.