AnythingLLM

Open-source desktop and Docker app for private chat over your own documents

AI Chat AssistantsFree planOverseas★★★★☆ 4.0

Visit AnythingLLM official site ↗ View similar tools Compare alternatives →

What is AnythingLLM?

AnythingLLM is an open-source application from Mintplex Labs for running private chat and retrieval over your own documents. You install it as a desktop app or run it with Docker, feed in PDFs, spreadsheets, code and web pages, and organise them into separate workspaces that each keep their own vector store and settings. It is model-agnostic: point it at OpenAI, Anthropic, Gemini, Groq or a local model server and it uses whichever you choose. The differentiator is packaging, since ingestion, embedding, vector storage and the chat interface all ship as one system you host yourself, so nothing has to leave your machines.

Last updated: 2026-09-20. This site only provides an index; for exact features, pricing, and licensing, see the official website.

Key features

Workspaces that isolate documents, vector data and chat history per project
Ingests PDFs, DOCX, spreadsheets, code files, URLs and plain text
Connects to OpenAI, Anthropic, Gemini, Groq or local Ollama and LM Studio
Built-in vector database options plus support for external vector stores
Agent skills that let the model browse the web, run code and call tools
Desktop app, Docker image and multi-user deployment from one codebase

Pros & cons

Strengths

Runs entirely on your own hardware with no data leaving your network
Model-agnostic, so switching providers does not mean rebuilding your setup
Free and open source with a Docker path for shared team use

Watch out for

Ingestion quality still varies with how messy the source documents are
Self-hosting asks for real comfort with Docker, ports and API keys
Large document collections want decent hardware and a proper vector store

Best for & use cases

private document chat, retrieval over internal files, self-hosted assistants, research notes

If you're comparing similar products, check the alternatives below, or browse all tools in the AI Chat Assistants category.

FAQ

Do I need an API key to use AnythingLLM?

For cloud models, yes, you supply keys for OpenAI, Anthropic, Gemini or another provider. It can also run fully offline against a local model server such as Ollama or LM Studio, in which case no external key is needed at all.

Is my data sent anywhere?

Only to whichever model provider you configure. Embeddings and chat can both be pointed at local models, and the vector store stays on your machine, so a fully offline setup is possible for sensitive material.

How is it different from a plain chat client?

A chat client answers from model knowledge. AnythingLLM adds document ingestion, embedding and a vector store per workspace, so answers can be grounded in your own files instead of general training data.