Use Hugging Face with NimbleBrain.
Connect Hugging Face and just ask — no workflow to build. Build, train and deploy state of the art models powered by the reference open source in…. The agent gets 40 actions, running under your own credentials.
Actions
Everything the agent can do in Hugging Face
40 actions
- Create ask access
- Create collection
- Create datasets tag
- Create discussions
- Create models tag
- Create papers index
- Create repo
- Create spaces tag
- Create webhook
- Delete datasets tag
- Delete discussions
- Delete spaces tag
- Get daily papers
- Get dataset info
- Get dataset rows
- Get dataset size
- Get datasets jwt
- Get datasets scan
- Get discussion
- Get discussions
- Get jobs hardware
- Get model info
- Get models compare
- Get models jwt
- Get models scan
- Get resolve
- Get settings mcp
- Get space info
- Get spaces compare
- Get spaces events
- Get spaces jwt
- Get spaces metrics
- Get spaces resolve
- Get spaces scan
- Get trending
- Get users avatar
- Get users overview
- Get users socials
- Get whoami
- List collections
Getting connected
Three steps, and Hugging Face is reachable.
-
Install it
Find Hugging Face in the directory inside your workspace and install it. No adapter for us to build first, no queue to join.
-
Connect the account
Sign in at the vendor, or hand an admin credential over once. Tokens stay with the workspace or the person who owns them — never pooled, never copied to us.
-
Ask for the work
The tools show up in front of the agent, under the same access wall as everything else. Describe the job in plain language, or put it on a schedule.
