Download
Running in a minute.
Python 3.10 or later is the only requirement. MAFDET installs llama.cpp and the models by itself, with a progress bar for every download. Current version: v0.4.1, release notes on GitHub.
With a double-click
- Download MAFDET: on GitHub, Code → Download ZIP, and unzip it (or
git clone). - Double-click
start_mafdet.bat. The first time it creates a Python environment in.venvand installs MAFDET, about a minute. Needs Python 3.10 or later. - After that, a double-click opens MAFDET in under a second. After an update it installs the new libraries by itself.
From PyPI
With Python from python.org, pip works as it is.
$ pip install mafdet $ mafdet
Runs on NVIDIA through CUDA (reference: RTX 2060, 6 GB), on AMD Radeon through Vulkan (in place, not measured yet), or on the CPU.
With a double-click
- Download MAFDET (Code → Download ZIP, or
git clone). - Double-click
start_mafdet.command. Python 3.10 or later: from python.org, orbrew install python@3.12. - If the Finder says it cannot verify the file: right-click, Open, then Open again. Only the first time.
From PyPI, with pipx
Homebrew's Python refuses pip install outside a virtual environment: pipx gives MAFDET an environment of its own.
$ brew install pipx $ pipx ensurepath # then open a new terminal $ pipx install mafdet $ mafdet
Apple Silicon (M1 or later) runs the models on the GPU through Metal, in the unified memory. Tested on an M2 Pro with 16 GB. Intel Macs run on the CPU.
Not tested yet
Linux has its llama.cpp builds and code paths in place (CUDA, Vulkan, CPU), but nobody has run MAFDET on it end to end yet. If you try it, an issue with the output of mafdet doctor is the most useful thing you can send.
From PyPI, with pipx
Recent Ubuntu and Debian stop pip install with externally-managed-environment: use pipx.
$ sudo apt install pipx $ pipx ensurepath # then open a new terminal $ pipx install mafdet $ mafdet
In a terminal
Also for the measurements in research/.
$ git clone https://github.com/TheLastOxygen/mafdet.git $ cd mafdet $ python -m venv .venv # Windows: py -3.12 -m venv .venv $ source .venv/bin/activate # Windows: .venv\Scripts\Activate.ps1 $ pip install -e . $ mafdet
Updating
$ pipx upgrade mafdet # or: pip install -U mafdet
The double-click scripts update their environment by themselves after a git pull.
Either way: MAFDET asks whether to open the window or the terminal app, looks at your machine, installs the right build of llama.cpp (CUDA, Vulkan, Metal or CPU) and downloads the model that runs best. No PyTorch, no manual downloads.
Requirements
Built for the hardware most people own.
Windows
An NVIDIA GPU with up-to-date drivers (tested target: RTX 2060, 6 GB), an AMD Radeon through Vulkan (in place, not measured yet), or just a CPU.
Tested
macOS
Apple Silicon (M1 or later): the models run on the GPU through Metal, in the unified memory. Tested on an M2 Pro with 16 GB, macOS 26.2: Qwen3-4B reads 593 tokens/s and writes 57. Intel Macs run on the CPU.
Tested
Linux
The llama.cpp builds and code paths are in place (CUDA, Vulkan, CPU), but nobody has run it end to end yet. Tell us how it went.
Not tested yet
Python
3.10 or later (tested with 3.12). No PyTorch: llama.cpp runs the models, transformers only provides the tokenizers.
Disk
About 0.7 GB for llama.cpp with CUDA (100 MB with Vulkan, 60 MB for the CPU), plus the models: from 0.6 GB (Qwen3-0.6B) to 2.3 GB for the 4B recommended on 6-8 GB cards.
Memory
16 GB of RAM recommended for CPU-only runs and for models larger than the VRAM. Models larger than the VRAM run partly on the CPU: it works, but slower, and less reasoning fits before Enter.
The first start
What happens when you open it.
- 01
Window or terminal? The launcher asks, and can remember your choice (
mafdet --askshows it again). - 02
Your machine. MAFDET looks at the GPU, the memory and the CPU, and installs the right build of llama.cpp: CUDA for an NVIDIA card, Vulkan for a Radeon, Metal on an Apple Silicon Mac, CPU otherwise.
- 03
The model. The catalog says what every model does on your hardware before you download it, and marks the one that runs best. Qwen3-4B is the pick on a 6-8 GB card.
- 04
Chat. Start typing: the status line shows what the model reads and thinks while you write. Every answer ends with its numbers.