Picture published by the creator on GitHub.
FreeToken: Run large open AI models on your gaming PC
Added · Updated
Made by FlashML-org
View profile on GitHub (opens in a new tab)Run big open-weight AI models on your own gaming PC and chat with them.
For developers and hobbyists who want local AI for coding agents.
What it does
FreeToken is an engine that serves Mixture-of-Experts models, large AI models built from many smaller expert parts, on personal hardware. It shares work across your GPU, CPU and memory, caches recently used experts, and offers OpenAI- and Anthropic-compatible connections so chat apps and coding agents can talk to it.
How to use it
Download the desktop app for Windows or Linux; it sets up the engine and gives you a window for running models, chatting and tuning. Developers can instead install the Python package, start a server with a model folder or Hugging Face id, then chat in the terminal shell.
Common use cases
- Point the server at a downloaded Qwen3.6-35B-A3B model folder and ask 'What is a Mixture-of-Experts model?' and a streamed text answer appears in reply.
- Run the launch command naming Claude Code or Codex and that coding agent is installed if missing, its settings are written, and it opens connected to your local server.
- On a laptop without a GPU, open the terminal chat shell with the address of a FreeToken server on another machine, then type a question and see the model's reply in your terminal.
Can I use this?
Requirements The desktop app handles engine setup for you. For the command line, create a Python environment with uv or pip, install the package with its accel extras, and have a CUDA 13 toolkit ready because GPU kernels compile on first use.
Caveat Nightly test builds change file names every night; if you use them, link to one specific build file, not the moving nightly tag.
Worth knowing Native GPU support is named for NVIDIA RTX 30, 40 and 50 series; AMD users follow a separate install guide. Licensed under Apache-2.0.
Why it's here
Not trending this week
Last seen on the GitHub trending list on 4 November 2026. It stays on BuildTube so you can still find it.
- 16,825 people have starred it on GitHub.
Numbers checked on 8 November 2026; not refreshed since.
Behind it
View setup guideBuildTube has not run or verified this project. Everything above is written from what the creator published.
Made this?
Quick questions
- What is FreeToken?
- FreeToken is an engine that serves Mixture-of-Experts models, large AI models built from many smaller expert parts, on personal hardware.
- What can I use FreeToken for?
- Point the server at a downloaded Qwen3.6-35B-A3B model folder and ask 'What is a Mixture-of-Experts model?' and a streamed text answer appears in reply; Run the launch command naming Claude Code or Codex and that coding agent is installed if missing, its settings are written, and it opens connected to your local server; On a laptop without a GPU, open the terminal chat shell with the address of a FreeToken server on another machine, then type a question and see the model's reply in your terminal
- Can I run FreeToken on my computer?
- For the command line, create a Python environment with uv or pip, install the package with its accel extras, and have a CUDA 13 toolkit ready because GPU kernels compile on first use. Native GPU support is named for NVIDIA RTX 30, 40 and 50 series; AMD users follow a separate install guide.
- What do I need to use FreeToken?
- Desktop app for Windows or Linux. The command-line route needs Linux x86_64, an NVIDIA GPU with driver r580+ (CUDA 13), and Python 3.10 or newer.
- What license does FreeToken use?
- Licensed under Apache-2.0.
More from this week
- kimi-k3-in-c Run a 2.78-trillion-parameter model on 8 GB of RAM
Get text answers from the huge Kimi K3 AI model using one CPU and no graphics card.
- agent-desktop Let AI agents read and click desktop apps
For developers building computer-use agents: agent-desktop lets an AI agent read and operate desktop apps on Mac, Windows or Linux from a command line.
- Auto-Ontology Give AI agents checked business meanings for your database
Help AI agents answer questions about company databases using reviewed definitions and SQL, with each answer showing its steps.
Explore by topic
Post this
Download share imageFor makers
Made this? Show it’s on BuildTube — add the badge to your README or embed the card on your site.
README badge (Markdown)
[](https://buildtube.app/p/gh-flashml-org-freetoken/)Embed card (HTML)
<iframe src="https://buildtube.app/embed/gh-flashml-org-freetoken/" width="560" height="240" style="border:0;max-width:100%" loading="lazy" title="FreeToken on BuildTube"></iframe>