plugin / local-modelsv0.1.0 stable

Discover, verify, and wire local MLX-optimized models for an Apple Silicon coding agent.

Query the Hugging Face Hub, match MLX models to the host's memory and installed runtimes, verify candidates, and emit configuration for the selected agent runtime.

/capabilities

One path from search to configuration

Built for macOS on Apple Silicon with Python 3.9 or newer.

01

Real sizing

Compare actual quantized download size with available host memory.

02

Runtime-aware

Detect supported local runtimes and produce the relevant target configuration.

03

Verify first

Test-generate with a candidate before recommending or wiring it.

/release

Release trust

Version, source, license, and verification state are explicit.

Version
0.1.0
Status
stable
License
MIT
Last verified
2026-07-16