IntegrationsHugging Face integration
Every model on the Hub.A shortlist in one ask.
Connect Hugging Face to Claude Code, Codex, Cursor or any MCP agent through CoreSpeed. Your agent searches models, datasets, Spaces and papers on the Hub, reads a repo's files and commits, and follows its discussions. It signs in with your own Hugging Face account, and writing, such as opening a discussion, is a separate call.
From a searchto a shortlist.
Requests you can make once Hugging Face is connected. The data is fictional.
- Hugging FaceModels · sentence similarity
- acme/embed-small-v2Apache-2.0 · 1.2M downloads
- acme/embed-baseMIT · 640k downloads
- lab/mini-embedApache-2.0 · 210k downloads
Search the Hub
Public repos, and private ones you can read
- Hugging FaceCommits · main
acme/embed-small-v2
- Nov 6 · Update the tokenizer config
- Oct 28 · Add ONNX weights
- Oct 2 · First release
Check a repo's history
Read-only
- Hugging Faceacme/embed-small-v2 · #14
- SStewartNov 7Does v2 change the max sequence length?
- MMiraNov 8Yes, it is 512 now. See the config change.
Follow the discussion
Commenting is a separate call
Name the job.Get the models that fit.
Your agent searches the Hub, opens the top candidates, and compares license, size and recent activity. You get a shortlist with the reasons.
Reads fromHugging Face
All connectors →Find three small English embedding models on Hugging Face with a permissive license. Compare downloads, size and their latest commits, and recommend one.
hugging-face__list_modelsok
24 matches, sorted by downloads
Keep your agent.Bring your work.
One CoreSpeed account, across compatible agents.
- 1
Set up CoreSpeed
Give your agent the setup line. It adds CoreSpeed and signs you in.
- 2
Connect Hugging Face
Authorize the account and the content you need.
- 3
Start with one task
Ask for a summary, a brief, or a draft. Review the answer.
Setup guidesClaude CodeCodexCursor