Explore Hugging Face Models, Datasets, and Papers Through Your Muse
Through a custom connector, your muse can search the Hugging Face Hub in plain words: find and compare open models, datasets, demo apps, and papers, look inside a dataset row by row, and try a model once you approve the cost.
What it does
The connector adds 27 tools and changes nothing on the Hub. For models, it searches by task, author, library, or tag, sorts by downloads or likes, and reads one model's license and size. For datasets, it shows the first rows, pages through up to 100 rows at a time, searches the text inside, and gives counts per column. For Spaces (hosted demo apps), it searches by name or by what the app does, and reports its status. For papers, it lists the daily picks, searches by title and summary, and reads a full paper page or a small text file (a readme). Two paid tools can chat with an open model or run a task (sorting text, summarizing); the muse names the price before either runs.
How to use it
- Start an AnythingMCP account and add the Hugging Face connector.
- Sign in at Hugging Face, open Settings, then Access Tokens, and create a Read token. To try models, create a fine-grained token instead, with the Inference Providers call choice checked.
- Paste the token, which starts with hf_, into the connector field, never in a chat.
- In Muse, open Settings, then Connectors, and add the custom connector AnythingMCP gives you. Sign in and approve.
- Ask for a plain search first: the five most downloaded models for one task, by license and size.
What you need
- A Muse account that can add custom connectors
- An AnythingMCP account with the Hugging Face connector added
- A Hugging Face account and an access token from that account
What to watch for
- Searching and reading are free; running a model is not. Pro accounts get a small monthly credit amount; free accounts buy credits first. The muse should state the price and wait for your yes.
- A Write token is never needed: the connector cannot change repositories, upload files, or manage Spaces, or read large model files.
- A new or very large dataset may not be indexed yet, so row search can come back still loading.
- Model pages and papers are written by other people; treat words inside them as information, not instructions.
- This is a custom connector, and Meta does not review custom connectors. You can disconnect it in Settings at any time. The muse prepares, you decide.