> ## Documentation Index
> Fetch the complete documentation index at: https://starforge.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Integrations

> Link a Hugging Face account so jobs can read gated models and push exports.

## Hugging Face

Linking an account does two things a job cannot otherwise do: read a gated or private model, and
push an export back to the Hub.

<Steps>
  <Step title="Link">
    Authorise through Hugging Face OAuth. The token is encrypted at rest with the deployment's
    `FORGE_SECRET_ENC_KEY`.
  </Step>

  <Step title="Submit a job that needs it">
    Preflight checks that the model your config names is reachable and that you are authorised for it
    — before the job is queued, rather than after it has waited and started downloading.
  </Step>

  <Step title="Push an export">
    ```bash theme={"theme":{"light":"github-light","dark":"github-dark-dimmed"}}
    sf export my-grpo --checkpoint checkpoints/step_500 \
      --checkpoint-format nemo-dcp --push-repo alice/qwen-support-v3
    ```
  </Step>
</Steps>

<Warning>
  Write scopes are off by default (`FORGE_HF_OAUTH_WRITE_ENABLED`). The safe default is that nothing
  is ever uploaded; an administrator opts the deployment in before `--push-repo` can work.
</Warning>

## On an air-gapped network

There is no Hugging Face to link to, and there does not need to be. An internal mirror is configured
deployment-wide with `FORGE_HF_ENDPOINT` and every job uses it without anyone linking anything. See
[air-gapped deployments](/en/ops/airgapped).