anything-llm

AnythingLLM is an all-in-one RAG platform that enables AI chat with any documents and data sources. Upload PDFs, Word files, web pages, and more to have AI understand and answer questions about their content.

• Multi-model support selecting from OpenAI, Anthropic, Ollama, LM Studio, and more
• Document RAG ingesting knowledge from PDFs, DOCX, TXT, web pages, and YouTube subtitles
• Workspaces for organizing documents and chat history by use case
• Agent capabilities with web search, code execution, and file operation tool calling
• Multiple vector DB support for Chroma, Pinecone, Weaviate, and LanceDB
• User management with multi-user support, permissions, and chat sharing

Built with TypeScript and React, AnythingLLM supports Docker self-hosting and offers a desktop app. It is ideal for building private AI assistants grounded in internal documents as a ChatGPT alternative.