REVIEW · LOCAL DOCUMENT CHAT
Verdict: AnythingLLM is one of the best free ways to run a private AI assistant on a Mac: open source, no account, models and documents kept on your machine, plus document chat, meeting notes and local file tools. It is a general AI workspace; if your only goal is finding a lost file fast, a dedicated search app is simpler.
1 November 2026 · By Ishtiaque Hossain · 4 minute read
Looking for a file you can describe but not name? Try Shlexball. The first 30 searches are free and need no account.
macOS 15 or later. Apple silicon. 16 GB memory. On a phone? Email yourself the link.
This review is based on AnythingLLM's own pages, read on 14 October 2026. Every claim links to the page it came from. We are the makers of Shlexball, so we say plainly where AnythingLLM does better.
| Price | Free download; MIT-licensed open source |
| Runs on | macOS (Apple silicon and Intel), Windows and Linux; Android app |
| Account | None needed: no sign-up, no email, no credit card |
| Where your data goes | Models, documents and chat history stay on your machine |
| Made by | Mintplex Labs |
| Best for | Anyone who wants a free, private AI assistant with document chat |
AnythingLLM is an all-in-one AI application: chat with AI models, work with documents through RAG, and run AI agents, without code or setup headaches. It is built by Mintplex Labs and its open-source community.
The desktop app runs entirely on your computer. Download one file, pick a model or let AnythingLLM recommend one for your hardware, and you are ready, with no cloud API key needed.
It is MIT licensed and open source, with more than 67 thousand stars on GitHub, and organizations can self-host a multi-user version with Docker on their own servers.
Document Knowledge turns every document you add into part of your assistant's knowledge base, with no data leaving your device. RAG, short for retrieval-augmented generation, means the app pulls the relevant passages from those documents to answer your question.
The desktop app also has File System Capabilities: AnythingLLM says it can read, write and search local files without you dragging them into the app.
Its Meeting Assistant records, transcribes and summarizes meetings on your own computer, with action items and full transcripts, and without a bot joining the call. Version 1.17.0 rebuilt it with smaller models and better speaker labels.
There is also on-device dictation, web search and scraping, background jobs and custom agent skills.
Local is the default. AnythingLLM says your models, documents and chat history stay on your machine and nothing phones home, and the app needs no account.
Some features reach the internet by design. Web search, powered by You.com, is on from your first chat, so a question sent to web search does leave your Mac.
Under the hood, the built-in engine now runs llama.cpp managed by AMD's Lemonade, after AnythingLLM moved away from Ollama in October 2026.
The desktop app is free to download, with native builds for Apple silicon and Intel Macs, Windows on x64 and ARM, and Linux.
Because models run on your own hardware, speed depends on your Mac. AnythingLLM picks a model suited to your hardware if you let it.
There is also an Android app that runs small models on the phone and can pair with your desktop to reach larger models and your workspaces.
Partly. Its File System Capabilities can search local files, and Document Knowledge answers questions from documents you add. It is built as a general assistant, though, so finding a file is one task among many rather than the whole design.
Keep in mind that the same file tools can also write files, which is powerful for agents and worth understanding before you point one at important folders.
Shlexball does one job: it finds files on your Mac from a plain-English description of what you remember, searching file and folder names and the text inside files, including PDFs and Word documents.
It is read-only, so it never changes, moves or deletes files, and its language model runs on your Mac. The first 30 searches are free and need no account.
No uploading: Shlexball searches the whole Mac and opens the file. 30 free searches.
▶ DOWNLOAD SHLEXBALLYes. The desktop app is a free download, and AnythingLLM is MIT licensed open source.
By default, your models, documents and chat history stay on your machine and the app needs no account. Web search, powered by You.com, does send your web questions out.
Yes. There are native downloads for Apple silicon and Intel Macs, as well as Windows and Linux.
Its File System Capabilities can read, write and search local files without dragging them into the app, and Document Knowledge answers questions from documents you add.