×
A Hugging Face Indexer
Face-Hugger is a search engine for Hugging Face repositories. Why? Because their built-in search is... let's go with "charmingly inadequate." If you've ever tried to find a specific LoRA or model file on their site, you know the pain of scrolling through thousands of results with no way to filter by file type, model architecture, or verified training data. So here we are, fixing their search for them.
The Profile popup includes a Customize tab where you can recolor the verified type and sub-type badges that appear on search results and metadata popups. Each model family has its own default color, and you can:
Custom colors are saved alongside your favorites and included when you export/import your profile.
⚠ Note: All settings are stored in your browser's local storage. If you clear your browser data or cache, your favorites will be deleted. To avoid losing them, use the Export button in the Profile popup to save a backup file, and Import to restore it later.
Some repos require you to accept terms or request access before downloading. You'll see two types:
Currently, certain model files are excluded from analysis to keep storage costs down. Some files can't be successfully analyzed due to timeouts, access restrictions, or corruption. Depending on the file format, different amounts of metadata are embedded. Safetensors files often contain rich training data (base model, trigger words, network dimensions), while other files may have little to none.
If this site gets popular enough or receives enough donations, we can expand what gets analyzed and make more metadata filterable. Currently file searching mainly works with .safetensors and .gguf files. In the future, config files and README contents can also be indexed and searchable.
This project is still a work in progress, and we'll continue to work on it and add new features. If you find any bugs or have suggestions please join the Discord server!
Join the Discord