react-native-vision-rtc

A React Native library for streaming high-performance camera frames: Vision Camera, ARKit, OpenCV, or ML, over WebRTC with zero-copy GPU transfer.

• Plugin-based: Add filters, ML, or AR processors only as needed.
• Type-safe, cross-platform, and unopinionated about signaling.
• No per-project native code required.

🚧 This library is a work in progress. APIs, behaviors, and native implementations may change without notice.

Features

• Zero-copy streaming of native GPU frames from Vision Camera and other sources.
• Composable processor pipeline for filters, ML, and AR (via plugins).
• Live track management: pause/resume, replace, adjust constraints, etc.
• Minimal core: install only the plugins you need.

Roadmap

• First release will support Vision Camera as a source.
• Skia GPU compositor and other processors coming as plugins.

Made with create-react-native-library