Hub Stats

Hub Stats

Shows growth, downloads, licenses, model trees, trending repos and papers across the Hugging Face Hub as one static page with every number precomputed. For people tracking Hub activity.

Made by huggingface on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: Hub Stats is a single static web page of statistics about the Hugging Face Hub. It reports how the Hub is growing, what gets downloaded, which licenses are used, how model trees branch, and which repos and papers are trending. The numbers come from the cfahlgren1/hub-stats dataset and are worked out ahead of time, so your browser runs no queries.

Who it suits: It suits people who follow what is happening on the Hub and want the numbers on one page.

Industry | Who uses it | How they use it | Example
AI research | Researchers | Track growth and trends across the Hub | Look at which licenses are common
Open source | Maintainers | See how downloads and trending repos move | Check where a repo type is heading
Journalism | Writers on AI | Find figures for a story | Pull the raw numbers from stats.json
Data engineering | Data engineers | Study a static build fed by a dataset | Read build/build.py

Skip it if you want to ask your own questions of the data, since the page shows fixed, precomputed numbers.

Where to use it:

• A web browser, on the Space's page at Hugging Face
• A local copy of the files on Windows, macOS or Linux, opened from index.html

Use case | Who | What they get back
See Hub-wide numbers | Researchers | Charts and tables on one page
Get the raw figures | Analysts | A stats.json file published beside the page
Rebuild the page | Data engineers | A fresh index.html from the dataset

How to set it up: The online page needs nothing. Python is needed only to fetch the files for a local copy. Rebuilding the page needs uv.

• A web browser
• Python and pip, only to fetch a local copy
• The uv tool, only to rebuild the page

1. Install the Hugging Face hub library.
2. Download the Space's files into a folder named hub-stats.
3. Move into that folder.
4. Load index.html in a web browser.

pip install -U huggingface_hub
hf download huggingface/hub-stats --repo-type space --local-dir hub-stats
cd hub-stats

How to use it: Open the page and read the sections. To rebuild the page yourself, run the build script with uv.

uv run build/build.py              # build locally into dist/
uv run build/build.py --publish    # build and push index.html + stats.json to this Space

The second line pushes to the Space, so it only works for someone who is allowed to publish there.

Example: You open the page to see which licenses dominate. You then open stats.json to take the underlying numbers for your own spreadsheet.

For newcomers and experts: Open the page and scroll. Everything is already calculated, so nothing loads slowly and nothing needs a login.

The page is one index.html of roughly 75 KB with every figure inlined. The build script uses DuckDB to read only the columns it needs from the dataset's parquet files, works out every statistic in about 25 seconds, and renders a template into the page. A scheduled Hugging Face job runs it daily, shortly after the dataset refreshes. The trade-off is that nothing on the page is live between runs.

The dataset refreshes
A scheduled job runs build.py
DuckDB computes every stat
A template becomes index.html
stats.json is published beside it

Good to know:

• Freshness: The page is rebuilt once a day, so it lags the Hub by up to that long.
• Page size: About 75 KB, with no downloads or queries in the browser.
• License: apache-2.
• Cost to run: none named for readers. The scheduled job uses a Hugging Face token as a secret.
• Data it reaches: A static page reads nothing outside itself. The build reads the cfahlgren1/hub-stats dataset.

On Hugging Face, 13 November 2026

• Likes: 65
• Built with: static
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/huggingface/hub-stats
• Files: https://huggingface.co/spaces/huggingface/hub-stats/tree/main

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.