BitRun
Zero-compile runner and terminal dashboard for 1-bit ternary LLMs on CPU.
Tested on macOS and Linux
Use it
macOS arm64
download install.sh from GitHub
curl -fsSL https://raw.githubusercontent.com/IshekKhal/bitrun/main/install.sh -o /tmp/install.sh && wc -l /tmp/install.sh && head -n 200 /tmp/install.shrun install.sh to extract bitrun binaries to ~/.bitrun/bin
sh /tmp/install.sh
Check that it works
export PATH="$HOME/.bitrun/bin:$PATH" && bitrun --version 2>&1The output should contain “bitrun version 0.2.0-beta”.
First thing to try
export PATH="$HOME/.bitrun/bin:$PATH" && bitrun hardware 2>&1Run bitrun hardware to see what your machine offers; you may first need to add $HOME/.bitrun/bin to your PATH.
Linux x86_64
Download install.sh
curl -fsSL https://raw.githubusercontent.com/IshekKhal/bitrun/main/install.sh -o /tmp/install.shRun install script
sh /tmp/install.sh
Check that it works
export PATH="$HOME/.bitrun/bin:$PATH" && bitrun --versionThe output should contain “bitrun version 0.2.0-beta”.
First thing to try
export PATH="$HOME/.bitrun/bin:$PATH" && bitrun hardwareRun bitrun hardware to see what your machine offers; you may first need to add $HOME/.bitrun/bin to your PATH.
What it does
- Run 1-bit ternary language models on your CPUTested
How to use it
- Download a model with
bitrun pull <model>, for examplebitrun pull bitnet-70m. - Run
bitrun bench <model> --tokens <number>to generate tokens and see the speed.
- Download a model with
- Serve a model through a local OpenAI-compatible APITested
How to use it
- Start the server with
bitrun serve <model> --port <port>. - List models with
curl -s http://127.0.0.1:<port>/v1/models. - Send chat requests to
http://127.0.0.1:<port>/v1/chat/completionsas JSON with a model name and messages.
- Start the server with
- Download 1-bit models from Hugging FaceNot tested
- Open a full-screen dashboard in the terminalNot tested
How far pickfit got
macOS arm64
- Installs
- Starts
- 1 of 4 features work
- Works offline
Linux x86_64
- Installs
- Starts
- 2 of 4 features work
- Works offline
Other systems
Windows: the project says it runs there; pickfit hasn’t tried it.
Good for
- Pull a 1-bit model directly from Hugging Face
- Launch full-screen dashboard with live hardware telemetry
- Serve a model via local OpenAI-compatible REST server
How pickfit tested it
Each run starts from a clean machine and follows the steps above. How pickfit verifies
macOS arm64
Its main function works
- Replayed from scratch
- Works offline
- Checked on 1 Nov 2026
- Version 0.2.0-beta
Differences the test machine hit
- README: bitrun --help shows usage
In the test: bitrun --help prints help but exits 1, not 0 (bitrun --version exits 0)
Linux x86_64
It produces model output
- Replayed from scratch
- Works offline
- Checked on 1 Nov 2026
- Version 0.2.0-beta
Differences the test machine hit
- README: curl -fsSL ... | sh adds bitrun to PATH
In the test: Installer appends to ~/.bashrc; current shell needs export PATH="$HOME/.bitrun/bin:$PATH" - README: bitrun cockpit <model> launches dashboard
In the test: Requires /dev/tty; fails in headless sandbox with 'could not open a new TTY'