chroma

To build a local vector store for semantic search and RAG workflows, you can use this Skill to store high-dimensional embeddings and metadata. It provides a simple API to create collections, insert documents, and run queries from notebooks to production environments.

Yes, you can filter vector search results using metadata in an embedding database. This Skill supports metadata filtering alongside vector and full-text search, enabling structured document retrieval for your AI applications.

Yes, the self-hosted vector store works with LangChain and LlamaIndex integrations. It supports a server mode for multi-user access, allowing seamless integration with these frameworks for your LLM workflows.

The best way to manage collections in a self-hosted embedding database is using a simple 4-function API. This Skill provides functions to create, insert, query, and manage collections with persistent storage from local development to production.

You need server mode for your vector store when you require multi-user access to your embeddings. This Skill supports server mode, enabling multiple users to query the same collections and integrate with external frameworks.