blakeblackshear/frigate

Frigate is a network video recorder and AI-powered surveillance system that processes live video streams to detect objects, faces, and motion. It features a modular architecture supporting diverse hardware accelerators (GPU, NPU, TPU) and integrates generative AI for tasks like semantic search, object description, and audio transcription. The system provides a comprehensive web interface for managing cameras, reviewing events, and configuring complex detection and post-processing pipelines.

This period was defined by a comprehensive modernization of the Frigate codebase, shifting from a monolithic structure to a modular, type-safe architecture using FastAPI and Pydantic. The work focused on replacing legacy components with robust, scalable systems, such as migrating to SQLite-vec for embeddings and ZMQ for inter-process communication. Additionally, the web frontend was completely rebuilt with React and TypeScript to support new features like semantic search, user management, and hardware acceleration support.

84 changes

This period focused on integrating Generative AI capabilities across the platform, introducing a modular post-processing and real-time processing architecture for tasks like transcription, face recognition, and object description. Simultaneously, the web interface received a comprehensive overhaul of the settings and configuration forms, replacing static definitions with a dynamic, schema-driven system that supports complex UI components and deep internationalization. The release also expanded AI provider support and added background job infrastructure for long-running tasks.

38 changes

blakeblackshear/frigate was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point the surveyor at a repository you know and see whether you agree with it.