Project launch and benchmark notes

[PetoVeritas]

Sharing the initial public launch of MLX-TurboQuant-Service.

This project is a local macOS service that runs Gemma 4 through Apple’s MLX stack and exposes it as a stable provider boundary for apps like OpenClaw.

Current practical focus:

• reliable local serving
• MLX-based Gemma 4 integration
• real agent-style benchmark and memory comparisons

Repo:
https://github.com/PetoVeritas/MLX-TurboQuant-Service