ax-engine 6.5.1

Python bindings for AX Engine

AX Engine

High-performance local inference engine for Apple Silicon — Python bindings.

Installation

This package README describes the current 6.4.x Python package. If your package index only shows older ax-engine versions, those wheels may not expose the top-level ax-engine CLI commands shown below.

Python (pip)

python3 -m pip install "ax-engine[download]>=6.4.1,<7"

Requires macOS 26+, Apple Silicon (M2 Max or newer), Python 3.10+. The current macOS arm64 wheel includes the ax-engine orchestration CLI plus bundled ax-engine-server and ax-engine-bench binaries.

Verify the installed command surface:

ax-engine doctor
ax-engine-server --help

Command-line tools (Homebrew)

Homebrew is the native binary channel for tagged macOS arm64 releases, but the tap may lag behind the Python package. Check the formula version first:

brew info defai-digital/ax-engine/ax-engine
brew install defai-digital/ax-engine/ax-engine

Use Homebrew for the ax-engine orchestration CLI only when the formula is 6.4.1 or newer. Then verify:

ax-engine doctor
ax-engine-server --help
ax-engine-bench doctor

Quick start

import ax_engine

session = ax_engine.Session(mlx=True, mlx_model_artifacts_dir="/path/to/model")
result = session.generate([token_id, ...], max_output_tokens=128)
print(result.output_tokens)

Or use the OpenAI-compatible shim:

python -m ax_engine.openai_server \
    --model-id my-model \
    --mlx-model-artifacts-dir /path/to/model \
    --tokenizer /path/to/tokenizer.json \
    --port 8080

Then point any OpenAI client at http://127.0.0.1:8080.

Optional dependencies

python3 -m pip install "ax-engine[openai]>=6.4.1,<7"     # FastAPI + uvicorn
python3 -m pip install "ax-engine[multimodal]>=6.4.1,<7" # image/audio helpers

Requirements

• macOS 26 (Tahoe) or later
• Apple Silicon — M2 Max / M2 Ultra / M3 / M4 family (32 GB RAM minimum)
• Python 3.10+

Source

github.com/defai-digital/ax-engine