Skip to content

Hugging Face Spaces

Build, deploy, and debug ML demo apps on Hugging Face Spaces, including ZeroGPU setup.

Stars
11,253
Licence
Apache-2.0
Files
15 files

What it does

Once installed, the agent knows how to create and maintain apps on Hugging Face Spaces, which are hosted git-backed projects for machine-learning demos. It guides choosing between Gradio, Docker, and Static setups, picks hardware such as the free-for-visitors ZeroGPU option or paid GPUs, and uses the hf command-line tool to create and configure a Space. It also helps debug Spaces that fail to build or run.

Great for

  • Developers who want to publish a machine-learning model as a public web demo
  • People porting an existing Gradio or PyTorch app onto ZeroGPU
  • Users fixing a Space that will not build or start

Skip it if

  • Users who need a guaranteed paid GPU without any billing: the skill notes that dedicated hardware requires a paid plan and the creator's billing to be enabled
  • Sensitive or private production workloads, since the skill centers on public hosting and visibility settings that the user must choose carefully

Try asking

  • “Create a Gradio Space on ZeroGPU for my image classification model”
  • “Port this PyTorch demo to run with @spaces.GPU on Hugging Face”
  • “My Space fails to build with a requirements error, help me debug it”
  • “Find an existing Space for a text-to-speech model I can adapt”
Step-by-step guide: how to use Hugging Face SpacesInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 15 original files, the Apache-2.0 licence and an attribution note.

    huggingface-spaces.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip huggingface-spaces.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip huggingface-spaces.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Let the agent search existing Spaces first with hf spaces search, so it can reuse a working app.
  • Keep access tokens scoped to write only what is needed and never paste them into public places; pass secrets via --secrets rather than plain env values.
  • Check your account's canPay and isPro status before choosing dedicated hardware to avoid unexpected hourly charges.

What's inside

  • README.md
  • references/3d-cuda-extensions.md
  • references/3d-generation.md
  • references/3d-gsplat.md
  • references/3d-models.md
  • references/3d-outputs.md
  • references/buckets.md
  • references/debugging.md
  • references/gradio.md
  • references/grants.md
  • references/inference-providers.md
  • references/known-errors.md
  • references/requirements.md
  • references/zerogpu.md
  • SKILL.md

Redistributed unmodified from huggingface/skills under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.