AnythingLLM

AnythingLLM is the all-in-one AI application for teams and individuals who want to chat with their documents, build RAG pipelines, and run AI agents — all on their own infrastructure, with any LLM backend.

**Chat with your documents**: Upload PDFs, Word docs, spreadsheets, websites, or raw text. AnythingLLM chunks, embeds, and stores them in a built-in vector database. Ask questions and get answers grounded in your actual content — not generic LLM responses. This makes it ideal for legal docs, internal wikis, codebases, research papers, and customer support knowledge bases.

**Any LLM, any backend**: AnythingLLM connects to virtually any LLM provider: local models via LM Studio, Ollama, or LocalAI; cloud providers like OpenAI, Anthropic, Groq, and Azure; or custom API endpoints. Switch backends without losing your workspaces or data.

**Team workspaces and permissions**: The cloud or self-hosted server version supports multiple users, with workspace-level access controls. Different teams can have isolated document collections and model configurations within a single deployment.

**Built-in AI agents**: Beyond simple RAG, AnythingLLM supports AI agents that can browse the web, execute code, use tools, and chain actions across multiple steps — all configurable through the UI without writing code.

**Key features**: - Self-hosted (Docker) or managed cloud deployment - Built-in vector database (no Pinecone or Weaviate required) - Multi-document workspaces with access controls - OpenAI-compatible API for programmatic access - No-code setup — get running in minutes with Docker - Free and open-source (MIT licensed on GitHub at Mintplex-Labs/anything-llm)

AnythingLLM is the practical choice for organizations that want AI on their data without sending anything to external servers.