changelog.

• Added update manager! You can now automatically check and update modelhub ota.

This one's packed with loads of features!

Features

• Settings tab. Added a third tab to the menu (Local · Explore · Settings) for app preferences.
• Source toggles. Settings tab now lets you pick which sources show up under Local — Hugging Face on by default, LM Studio opt-in. Choice is persisted across launches.
• Click to open on Hugging Face. Clicking any row in the Explore tab opens that model's Hugging Face page in your browser. The download button still works on its own.
• Instant Explore landing. The default ("All Models" / "Fits This Mac") list is now cached in memory. Reopening the menu paints results instantly instead of flashing a loading state.
• "Runs on this Mac" toggle. Filters for Apple Silicon-runnable formats (MLX, GGUF, llama.cpp) and checks that the model fits in your Mac's unified memory.

Polishes and fixes

• Now a small popover appears to point at the menubar icon to denote that modelhub lives here
• Fixes and polishes to long text marquee and type/size columns
• State fix on downloading a model: now appears immediately in local tab

Browse and download HuggingFace models, in your menu bar.

The new Explore tab puts HuggingFace search, author info, model size, and one-click download (with live progress + pause/resume) right next to your local models. Downloads use the standard ~/.cache/huggingface/hub layout, so transformers, mlx-lm, llama.cpp, and the HF CLI all see them as native cache entries.

First release 🎉

Grab the Modelhub.dmg from the assets below, mount it, and drag Modelhub into your Applications folder. That's it! If anything breaks, report an issue!