AnythingLLM is an open-source Windows app by Mintplex Labs that combines AI chat, document workspaces and agents, using local or cloud models. Version 1.16.2.
Rate AnythingLLM
Average 0.0 · ratings: 0
Thank you! Your rating is saved; you can change it. Could not save the rating. Please try again later. Too many ratings from your network today. Please try again tomorrow. The average appears after 5 ratings.
About AnythingLLM
AnythingLLM 1.16.2 for Windows is an open-source desktop app from Mintplex Labs that puts chat, document search and AI agents into one workspace-based interface. It can run a model locally with its built-in runner, connect to Ollama or LM Studio on the same PC, or use an API provider in the cloud. The download on this page is the official installer from the vendor's CDN; version 1.16.2 was released on 22 September 2026.
How AnythingLLM is organised
The central idea is the workspace. Each workspace has its own documents, its own chat history and its own settings, including which model answers. Drop PDFs, text files or web pages into a workspace and AnythingLLM splits and indexes them, so the model can answer from their content — the technique usually called RAG, retrieval-augmented generation. Agents go a step further: they can use tools such as web browsing or file operations to complete a task in several steps. For servers and teams there is a Docker edition, while the desktop app described here is meant for one person on one computer.
Where do the models come from?
This is where AnythingLLM differs from most local AI apps. You are not tied to one engine:
- the built-in local runner downloads and runs a model for you;
- an existing Ollama or LM Studio installation can serve as the back end;
- API providers can be added for cloud models when you need a larger one.
With a local back end, your documents and questions never leave the machine. With a cloud provider, they are sent to that provider under its terms.
System requirements
AnythingLLM runs on Windows 10 or later and is also available for macOS and Linux. The vendor recommends 16 GB of RAM. A GPU is optional; NVIDIA cards through CUDA, AMD graphics and NPUs can speed up local models. The installer, AnythingLLMDesktop.exe, is about 390 MB. Details are on the system requirements page of the documentation.
AnythingLLM compared with other local AI apps
| App | Main focus | Open source |
|---|---|---|
| AnythingLLM | workspaces with documents and agents, many model back ends | yes (MIT) |
| Ollama | engine with a command line and a local API | yes |
| LM Studio | model browser and chat window | no |
| Jan | chat app for local and cloud models | yes |
| GPT4All | chat with local documents (LocalDocs) | yes |
Many people combine two of them: Ollama or LM Studio as the engine, AnythingLLM as the place where documents and conversations live.
Licence
AnythingLLM is free and released under the MIT licence; the source code is in the Mintplex-Labs/anything-llm repository. Models keep their own licences.
Pros and cons
- Pro: documents, agents and chat in one app.
- Pro: works with its own runner, with Ollama or LM Studio, or with cloud APIs.
- Pro: open source under MIT.
- Con: more settings to understand than a plain chat app.
- Con: comfortable use needs plenty of RAM.