AnythingLLM
All-in-one AI app to chat with your documents
AnythingLLM is a full-stack, all-in-one desktop and Docker AI application for chatting with your documents. With 54K+ GitHub stars, it's the most flexible way to build a private ChatGPT-like experience. Connect any LLM, any vector database, and any document source. No coding required—just upload your files and start asking questions.
Why AnythingLLM?
Teams have documents scattered across drives, wikis, and databases. Extracting insights requires reading through everything manually. ChatGPT can't access your private files, and building a custom RAG system requires weeks of development. You need a simple way to make your documents searchable and conversational without sending data to third parties.
How It Works
AnythingLLM provides a complete document chat solution out of the box. Upload PDFs, Word docs, websites, or any text—it automatically chunks, embeds, and indexes them. Connect your preferred LLM (local or cloud) and start asking questions. Workspaces keep documents organized, and the multi-user system lets your entire team collaborate securely.
What Is AnythingLLM?
AnythingLLM is a private, all-in-one AI document chat application. It supports 30+ LLM providers including Ollama, OpenAI, and Anthropic. Upload any document type, organize into workspaces, and chat naturally. Features include AI agents, custom embeddable widgets, and a full developer API.
Key Benefits
Why teams choose AnythingLLM
Chat With Any Document
Upload PDFs, Word docs, websites, or plain text. Ask questions in natural language.
Any LLM Provider
Connect to Ollama, OpenAI, Anthropic, or 30+ other providers. Switch anytime.
Workspace Organization
Keep documents in separate workspaces. Each workspace has its own context.
AI Agents
Build custom agents that can browse the web, write code, and use tools.
Embeddable Widget
Add a chat widget to your website. Your customers can chat with your docs.
Full Privacy
Run locally on your machine. Your documents never leave your infrastructure.
Features
Everything you need to build with AnythingLLM
Multi-User Support
Role-based access with admin, manager, and user permissions.
Custom Embeddings
Choose your embedding model for optimal search quality.
Vector Database Options
Support for LanceDB, Chroma, Pinecone, Qdrant, and more.
Developer API
Full REST API for building custom integrations.
Desktop App
Native desktop application for Mac, Windows, and Linux.
MCP Support
Model Context Protocol integration for extended tool capabilities.
Use Cases
What you can build with AnythingLLM
Technology Stack
Ready to deploy AnythingLLM?
Get started in minutes. Deploy on your own infrastructure at actual cloud cost. No markup, no vendor lock-in.