Chroma — векторная база для RAG и семантического поиска

★ 7.1 · data

chroma is a Claude Code skill that integrates the open-source Chroma vector database for building RAG pipelines and semantic search over documents. It depends on chromadb and sentence-transformers, and covers the full set of core operations: creating collections, adding documents with metadata, similarity queries, metadata filtering with logical operators, updates, deletes, and persistent storage via PersistentClient. Multiple embedding backends are supported out of the box — the default all-MiniLM-L6-v2 model, OpenAI text-embedding-3-small, HuggingFace models, and custom embedding functions. The skill suits developers who need a self-hosted, Apache 2.0-licensed vector store for prototyping LLM applications locally without managed cloud infrastructure.