Skip to content
ohmyhf

The Hugging Face desktop client that should have existed.

An unofficial, open-source app for browsing, downloading and managing the Hub — fast, local, private.

Apache-2.0 · macOS / Windows / Linux · No telemetry · English / 中文

Twenty tabs. Endless reloads. A 300 GB cache folder you're afraid to open.

ohmyhf fixes this.

01/Browse

Triple-pane browsing. Zero reloads.

Sidebar, list, and live preview stay put while you switch Models, Datasets, Spaces, and Papers. Long lists stay smooth with virtualization; the right pane renders model cards, file trees, and metadata instantly — no new tabs, no full-page reloads.

02/Command

Cmd+K. Keyboard-first Hub search.

One palette for search, filters, and sort. Narrow by task, library, license, or parameter count; rank by trending, likes, or downloads. Stay in flow without hunting through menus.

03/Downloads

A download manager the browser can't match.

Queue jobs, run them in parallel, pause and resume, cap speed, and verify sha256 on finish — with system notifications when you're done. Files land in the standard Hugging Face cache layout, so transformers and the CLI just work.

04/Cache

See your local cache. Reclaim the gigabytes.

Scan your Hugging Face cache directory, inspect disk use per repo, and clean old revisions in one click. The web can't do this — ohmyhf can save you hundreds of gigabytes in ~/.cache/huggingface.

Roadmap

On the way.

  • Coming soon

    Watch inbox

    Subscribe to users, orgs, and repos for updates and Daily Papers — with system notifications when something new lands.

  • Coming soon

    Model comparison

    Put 2–4 models side by side: parameters, license, and benchmarks in one glance so you can pick with confidence.

  • Coming soon

    Local AI toolchain

    Push GGUF into Ollama or LM Studio, send models to ComfyUI, and run an in-app inference playground — without leaving the desktop.

Principles

Built to earn trust.

  • Open source

    Apache-2.0 licensed. Read the code, fork it, audit it — the full repository is public on GitHub.

  • Privacy-first

    Everything stays local in SQLite. No telemetry. OAuth tokens are stored with OS-level encryption — never in plaintext.

  • Cross-platform

    Native builds for macOS, Windows, and Linux, with automatic updates so you stay current without hunting installers.

Get the app

Download ohmyhf

First release is on its way — watch the repo to get notified.

  • macOS

    dmg

    Coming soon
  • Windows

    nsis

    Coming soon
  • Linux

    AppImage / deb

    Coming soondeb

macOS builds are unsigned for now — see how to open them in the Opening unsigned macOS apps.

FAQ

Questions, answered.

Is this an official Hugging Face app?

No. ohmyhf is an unofficial open-source project. It is not affiliated with, associated with, or endorsed by Hugging Face, Inc. “Hugging Face” is a trademark of Hugging Face, Inc., used here for identification purposes only.

Is ohmyhf free?

Yes. The app is free and open source under the Apache-2.0 license. You can use it, study it, and contribute without paying a fee.

Are my data and tokens safe?

All app data stays on your machine in a local SQLite database. Tokens are stored with OS-level encryption, never in plaintext. There is no telemetry, and the source is public so you can audit what runs.

Which platforms are supported?

ohmyhf targets macOS, Windows, and Linux. Builds ship as dmg, NSIS installer, and AppImage / deb, with automatic updates once releases are published.

Is the cache compatible with huggingface-cli and transformers?

Yes. Downloads write to the standard Hugging Face cache layout. Models and datasets you fetch in ohmyhf work with huggingface-cli, transformers, and other tools that expect that layout.

How can I contribute?

We welcome issues, pull requests, and translations. Start with the contributing guide in the repository — it covers setup, coding standards, and how to propose changes.

CONTRIBUTING.md