beehiivYou published a model last month. Somewhere on the Hub it has a download count, and you check it when you remember, which is every couple of weeks if it’s lucky.
What the board shows from Hugging Face
Connect Hugging Face and the board can run the same searches you’d run on the Hub. Ask for your models sorted by downloads and each one shows its downloads and likes. Ask for what’s trending, or the most liked, and you get the models, datasets or Spaces everyone else is looking at.
Hugging Face’s server answers in words, so a widget shows its answer as a card, each model listed with its task, downloads and likes. The board reads it again while it’s open, so the card keeps up.
Put it beside your GitHub stars and the agent session training the next version, and the whole model, from code to downloads, is on one board.
Where the data comes from
Hugging Face runs an official MCP server for the Hub. The board connects to it without an account, for everything the Hub shares publicly. There’s no sign-in page and nothing to paste.
That also means a private model stays private. The board sees what the Hub shows everyone, and Hugging Face applies rate limits to anonymous use.
The board only runs tools that read. It never creates a repo, uploads a file or starts a job.
If a read fails, the widget keeps the last good answer and says what went wrong and when.
Why put Hugging Face on a board
The Hub is where models live. It isn’t where you spend your day. That runs through your code, your revenue and the agents doing the work, and the download count for the model you shipped is one more number you’d like to see without going to look.
On a board it sits beside your npm downloads, your stars and your agent sessions. Ask your agent to add it to the board you already have, or start from a blank canvas and build it around the model.