flamehaven01/Flamehaven-Filesearch

Supports hybrid search combining keyword matching with semantic vectors via the DSP v2.0 algorithm, which generates embeddings without ML framework dependencies. Built on FastAPI with PostgreSQL/pgvector for enterprise persistence, HNSW indexing for sub-millisecond vector retrieval, and integrates Gemini API for LLM-powered answers with optional multimodal image search capabilities. Includes OAuth2/OIDC authentication, per-API-key rate limiting and quota tracking, batch processing (1-100 queries), and audit logging for production deployments.

Used by 1 other package. Available on PyPI.