AnythingLLM
AnythingLLM is an MIT-licensed all-in-one AI application from Mintplex Labs for building a private ChatGPT. It chats with your documents and cites the sources, runs agents with a no-code builder and MCP tools, and works with any of 39 model providers, from local llama.cpp and Ollama to OpenAI and Anthropic. It ships as a desktop app for Mac, Windows and Linux, or as a Docker server with multiple users.
What does AnythingLLM do?
AnythingLLM is "the all-in-one AI application that lets you build a private, fully-featured ChatGPT", and its second line lists the jobs: "Chat with your docs, use AI Agents, hyper-configurable, multi-user, & no frustrating setup required." Documents such as PDF, TXT and DOCX are chunked into a vector database, LanceDB by default with 8 others listed, and answers come with source citations. The README lists 39 model providers, including any llama.cpp-compatible model, Ollama, LM Studio, LocalAI, OpenAI, Anthropic, AWS Bedrock, Google Gemini, OpenRouter, DeepSeek, Mistral, Groq and xAI.
Agents inside workspaces can browse the web and use tools, with a no-code agent builder and MCP compatibility. The README claims tool handling that reduces token use "by up to 80% per query", its own figure. It routes chats to providers by rules, keeps memories, and runs scheduled tasks. Voice uses the browser, Piper, OpenAI or ElevenLabs. The Docker version adds multi-user permissions and an embeddable chat widget. A developer API, a Chrome extension and a separate MIT mobile app are listed.
Key facts
- Licence: MIT, copyright Mintplex Labs, with no added conditions.
- 39 model providers listed, including llama.cpp-compatible models, Ollama, LM Studio, LocalAI, OpenAI, Anthropic, AWS Bedrock, Gemini, OpenRouter, DeepSeek, Mistral, Groq, xAI and LiteLLM.
- Document chat with source citations; LanceDB by default and 8 other vector databases; a built-in embedder.
- Agents with a no-code builder and MCP; memories; rule-based routing between models; scheduled tasks.
- Multi-user permissions and the embeddable chat widget are Docker only, per the README.
- Telemetry is on by default and turned off with DISABLE_TELEMETRY; the README says some outbound connections to its CDN and GitHub may still appear.
- Desktop app for Mac, Windows and Linux, or Docker with deploy guides for AWS, GCP, DigitalOcean, Render, Railway and others; a hosted instance is sold by Mintplex Labs.
- Latest tagged release when read: v1.16.1 on 2026-08-27; 66,151 stars on 2026-09-17.
What does it replace, and where does it fall short?
AnythingLLM replaces ChatGPT and Claude for someone whose main job is asking questions of their own documents, with citations, on a model they choose, local or hosted. It covers one person on a desktop and a team on a Docker server with the same code, and it is the document-chat pick on the open-source alternatives to ChatGPT page.
Where it falls short: multi-user support and the embeddable widget are Docker only, so the desktop app is single-user. Telemetry is on until you turn it off. The README does not describe image generation. The 80 percent token saving is the project's own claim. For a team platform with single sign-on and roles, Open WebUI or LibreChat.
How does AnythingLLM run?
Download the desktop app for Mac, Windows or Linux, or run the Docker image for a multi-user server; the README links deploy guides for AWS, GCP, DigitalOcean, Render, Railway, RepoCloud, Elestio, Northflank, Sealos and Easypanel, and a bare-metal guide. Choose a model provider in settings, local or hosted. Mintplex Labs sells a hosted instance through its own checkout page; the README states no price.
Who is AnythingLLM for?
A person or a small team that wants to chat with a pile of documents, with citations, on a model it controls, in a desktop app or on one server. A company that needs single sign-on and fine-grained roles should compare Open WebUI and LibreChat.
What limits does the README state?
From the README: multi-user support and the embeddable widget are "Docker version only"; telemetry collects anonymous usage information by default; with telemetry off, "you may still see outbound connections" to its CDN and GitHub.
Questions people ask
Is AnythingLLM open source?
Yes. The main repository is MIT with no added conditions, and the separate mobile app is also MIT by the README. Mintplex Labs also sells a hosted instance, which is optional; the desktop app and the Docker server are free to run.
Can AnythingLLM run fully offline?
Yes with local choices: any llama.cpp-compatible model, Ollama, LM Studio or LocalAI for the model, the built-in embedder, and LanceDB for storage. Turn telemetry off with DISABLE_TELEMETRY; the README notes that some outbound connections to its CDN and GitHub may still appear.
Does AnythingLLM support multiple users?
In the Docker version, yes, with permissions. The README marks multi-user support and the embeddable chat widget as Docker only, so the desktop app is for one person. For single sign-on and group roles, Open WebUI and LibreChat document more.
How does AnythingLLM compare with ChatGPT?
Its focus is your documents: upload them, ask questions, and get answers with citations, on any of 39 providers or a local model. ChatGPT offers OpenAI's models with search, images and voice as a service. The open-source alternatives to ChatGPT page places it as the document-chat pick.
Which vector database does AnythingLLM use?
LanceDB by default, which needs no setup, and the README lists 8 others as options. Embeddings come from a built-in embedder by default, or from a provider you choose. Chats cite the source documents the answer came from.
Sources
- AnythingLLM README and MIT LICENSE: github.com/Mintplex-Labs/anything-llm, read 2026-09-17.
Compared with the others
On the open-source alternatives to ChatGPT page, AnythingLLM is the pick for chatting with your documents. MIT, with document chat and source citations, agents with MCP, and 39 model providers including local llama.cpp and Ollama, as a desktop app or a Docker server.
Also on that page: Open WebUI for a private chatgpt for a whole team, LibreChat for every provider in one mit platform, Jan for a private chatgpt on one laptop.
On the open-source alternatives to Claude page, AnythingLLM is the pick for chatting with your documents. MIT, with document chat and source citations, agents with MCP, and 39 model providers including local llama.cpp and Ollama, as a desktop app or a Docker server.
Also on that page: LibreChat for claude plus agents, artifacts and a code sandbox, Cherry Studio for many providers and 300+ assistants in one desktop app, Chatbox for one app across desktop and phone.
On the open-source alternatives to Glean page, AnythingLLM is the pick for document chat with citations. MIT, chatting with your documents with source citations over 39 model providers, as a desktop app or a multi-user Docker server.
Also on that page: Onyx for company-wide search with agents, Open WebUI for documents plus roles and single sign-on, Khoj for notes, files and the web together.
More chat interfaces
A self-hosted chat platform for Ollama and OpenAI-compatible models with document search over 9 vector databases, web search, image generation, roles and single sign-on, under a BSD-3 licence with an added branding clause.
LobeHub, formerly LobeChat, is an agent workspace with an agent builder, agent groups, scheduled runs, editable memory and a library of over 10,000 tools and MCP plugins, under a community licence that needs a paid licence to distribute a modified version.
An AGPL-3.0 desktop client for Windows, macOS and Linux that connects OpenAI, Gemini, Anthropic and local Ollama or LM Studio models, with 300+ preset assistants, side-by-side model chats, file support and MCP servers.
An AGPL-3.0 desktop app for local models, formerly text-generation-webui, with five engines including llama.cpp and ExLlamaV3, an OpenAI and Anthropic-compatible API, tool calling, file chat, image generation and LoRA training, fully offline.
An Apache-2.0 desktop app from Menlo Research that downloads and runs open models from Hugging Face on your own computer, connects to OpenAI, Anthropic and others when you want, serves a local OpenAI-compatible API, and supports MCP.
Added September 17, 2026. Every claim above comes from the project's README, LICENSE or model card, read on September 17, 2026, or from the GitHub API on the date shown in the panel. Found an error? Write to reveneau@licheo.com and it is fixed in the next weekly pass. Repository: github.com/Mintplex-Labs/anything-llm.
Running one of these inside your own environment, with your own data and your own security rules, is the kind of work Reveneau does. Read how a forward deployed engagement works.