Chroma
Free tierOpen-source search infrastructure for AI — vector, full-text, regex, and metadata search at scale
Key strengths
Chroma is a serverless, multi-tenant vector database built on top of object storage (S3/GCS), enabling automatic data tiering between hot memory cache, warm SSD cache, and cold object storage. It supports sparse vector search (BM25, SPLADE), dense vector search, trigram/regex full-text search, and metadata filtering with faceted search. The query execution engine is written in Rust, and clients are available for Python, TypeScript/JavaScript, and Rust. Technical specs include p50 query latency of 20ms at 100k vectors (384 dimensions), write throughput of 30 MB/s per collection, and support for up to 1 million collections per database.
