AnythingLLM
Self-hosted RAG chat over your files, MIT-licensed
Overview
AnythingLLM is an open-source (MIT) workspace for chatting with your own documents — run it locally or in Docker, connect any LLM provider or a local Ollama model, and, in the Docker version, add multi-user workspaces with role-based access for team use.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 30 Sep 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- Full RAG feature set
- Bring your own LLM API key or local model
- MIT open source
- Private instance + custom subdomain
- Bring your own LLM API key
- Priority resources
- 72-hour support SLA
- SSO and RBAC
- Custom SLA
Use cases
What you can produce with AnythingLLM
- MIT-licensed, self-hostable (Docker or one-click desktop app)
- Connects 30+ LLM providers and local models via Ollama/LM Studio
- Multi-user workspaces with role-based access control (Docker version)
- Ingests PDFs, spreadsheets, GitHub repos, audio, and databases
- Built-in agent skills: web browsing and scraping, SQL, charts, file generation, Gmail and Outlook
- MCP client support, so agents can use tools from existing MCP servers
AnythingLLM is Mintplex Labs' open-source answer to "ChatGPT over my own files" — a desktop app or Docker container that ingests PDFs, spreadsheets, GitHub repos, and more into a private RAG chatbot, with a one-click installer and no account required to run it locally. It connects to 30-plus LLM providers plus local models via Ollama or LM Studio, and its agents can use tools from existing MCP servers. The desktop app and self-hosted Docker version are genuinely free; a managed cloud tier exists for teams that don't want to run their own infrastructure, starting at $50/month. It's a strong pick for developers and privacy-conscious teams who want document QA fully on-premise with no vendor lock-in. Non-technical users will find local-model setup fiddlier than a polished SaaS product, multi-user access needs the Docker version, and RAG answer quality depends heavily on which model and embedding settings you choose.
What we found
AnythingLLM comes from Mintplex Labs, a small Y Combinator team (Summer 2022) in Los Angeles led by Timothy Carambat, and its open-source repository — Mintplex-Labs/anything-llm on GitHub, MIT licensed — is among the more popular self-hosted AI tools. It ships often: version 1.16.2 landed on 22 September 2026. Early Hacker News threads praised the single-install desktop packaging as slick next to the DIY LangChain-plus-vector-database setups most self-hosted RAG projects require.
The feature list is broad for a project this size: more than 30 LLM providers (OpenAI, Anthropic, local Ollama and others), a choice of vector database (LanceDB is bundled by default, or plug in Pinecone, Qdrant and others), and ingestion of PDFs, Word documents, spreadsheets, HTML, audio, GitHub repositories and database connections. Built-in agent skills cover web browsing and scraping, SQL queries, document search, charts, file generation, and Gmail, Outlook and calendar. It also works as an MCP client, so its agents can use tools from existing MCP servers. Multi-user workspaces with role-based permissions need the Docker version rather than the desktop app, and there is now an open-source Android app.
Pricing splits cleanly: the desktop app and self-hosted Docker deployment are free with no account needed, which is the version most reviews focus on. For teams that don't want to run infrastructure, managed cloud starts at $50 a month for a private instance, then $99 for Pro with priority resources and a 72-hour support SLA, and custom Enterprise pricing for on-premise installs with SSO and role-based access.
Reviewers who use it daily describe it as a strong free alternative to a ChatGPT Plus subscription because you supply your own API key or run a fully local model. The honest caveats: RAG quality is only as good as your chunking and embedding settings, and non-technical users will find local-model setup — installing Ollama, pulling a model, wiring up the config — more involved than a polished SaaS tool. It is maintained by a very small team, which YC lists at four people, so support and roadmap pace depend on their bandwidth.
Compiled by RECATOOLS from public sources: vendor documentation and pricing, code repositories and published reporting.
About this listing
This entry was compiled from publicly available data including AnythingLLM's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with AnythingLLM unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to AnythingLLM directly →
Spotted something out of date? Suggest an update →
Related reading
From our guides
Guide · 17 min read
How to Self-Host AnythingLLM with Docker and Ollama (2026)
We self-hosted AnythingLLM 1.15.0 with Docker and Ollama, embedded a document ov...
Guide · 14 min read
LM Studio vs GPT4All vs AnythingLLM vs Open WebUI: The Best Local LLM Apps (2026)
We compare LM Studio, GPT4All, AnythingLLM and Open WebUI on licensing, platform...
Alternatives to AnythingLLM
More in Productivity
