Chroma is an open-source search and retrieval database designed for AI applications that need fast, scalable access to embeddings and related metadata. It provides vector similarity search, full-text search, metadata filtering, collection management, and multiple storage backends (ephemeral in-memory and persistent file-based). Chroma can be self-hosted or used via Chroma Cloud; it exposes flexible client types (HTTP, persistent, ephemeral, and cloud clients) and integrates into RAG pipelines and model context workflows. The project also provides MCP (Model Context Protocol) server implementations and tooling for remote access, packaging search scenarios, and hybrid retrieval approaches.