Connect Hugging Face to clariBI
Connect the Hugging Face Hub through clariBI's MCP catalog. No login needed for anonymous use, just live model, dataset, Space, and paper search next to your own analysis.
clariBI| Connection | MCP app. Public endpoint, no sign-in or key. |
| Access | Read-only. clariBI keeps only the Hugging Face tools that read. |
| Data | clariBI calls Hugging Face when you ask a question and syncs list data where the server offers it. |
| Plans | Trial, Starter, Professional and Enterprise. Not on Free or Lite. |
About Hugging Face
Hugging Face is a hub for open machine learning. It hosts models, datasets, and interactive Spaces, indexes machine-learning research papers, and publishes documentation for its libraries.
clariBI reaches the Hub through its hosted MCP server: semantic search over Spaces, a repository search across models, datasets, and Spaces, research-paper search, repository detail lookups, and documentation search. The endpoint is public and works anonymously, so no key is required, though Hugging Face rate-limits anonymous use.
Vendor site: huggingface.co
What clariBI can read
When you connect, clariBI lists the tools the Hugging Face MCP server advertises and keeps the read-only ones: Space search, repository search, paper search, repository detail, and documentation search. Tools that create, change or delete anything are filtered out.
Hugging Face is one of 86 apps in the clariBI MCP catalog. See the catalog
What you can chart
- Model, dataset, and Space search with org and tag filters
- Trending or recent repositories by sort order
- Research-paper search across the Hugging Face paper index
- Repository detail: downloads, likes, and metadata for a repo list
- Documentation search across Hugging Face and Gradio libraries
How to connect Hugging Face
Open Data Sources in clariBI
In clariBI, go to Data Sources > Add Data Source > MCP Servers tab. Find Hugging Face in the catalog and click Connect.
Connect, no sign-in needed
The Hugging Face endpoint is public and works anonymously, so there is no login screen and no key to paste. Clicking Connect registers the read-only Hub connector and you are done.
Ask questions, get dashboards
clariBI queries the Hub on demand. Ask "Which trending text-generation models from this org are largest?" and clariBI reads the public Hub and builds the comparison.
Questions you can ask
Asking in plain English uses AI credits. The AI decides which Hugging Face tools to call, totals are computed from the rows they return, and the AI writes the explanation.
- "Of the models we shortlisted, which have the most downloads and likes on the Hub?"
- "Which recent retrieval-augmented-generation papers reference models we already use?"
About the Hugging Face connector
Do I need a Hugging Face account or a key to connect?
No. The Hugging Face MCP endpoint is public and works anonymously, so clariBI sends no login or key by default. Anonymous use is rate-limited; Hugging Face documents setting an HF token for higher limits if you need them.
Can clariBI change anything on the Hub?
No. The connector reads public Hub data only: searching repositories, papers, and Spaces, reading repository detail, and searching documentation. It cannot create, edit, or delete a repository.
Which clariBI plans include Hugging Face?
Hugging Face needs the trial or a Starter, Professional or Enterprise plan. MCP apps are not part of the Free or Lite plans.
Try Hugging Face in clariBI
The trial lasts 14 days, includes 50 AI credits and needs no card. When it ends, your account moves to the Free plan and keeps its data and dashboards.