ENGINEERING

Maintain an approved-models registry in Notion as PRs clear license checks

When a PR that adds a Hugging Face model merges and the model's license passed policy, it upserts the model into a Notion approved-models registry with its license, task.

CategoryEngineering
Enginesim
Difficultyintermediate
Triggerevent
Steps5
Setup~15 min

How it runs

The automated pipeline, trigger to output.

  • TriggerPull request merged to mainGitHubGitHub
  • ActionExtract added HF model IDs from diffGitHubGitHub
  • ActionConfirm license and task from model cardHugging FaceHugging Face
  • LogicKeep only approved-license models
  • OutputUpsert model into Notion registryNotionNotion

What it does

Builds and keeps current a single source of truth for which Hugging Face models are cleared for use. Every time a license-compliant model lands on the main branch, it is recorded in a Notion database so the team can see exactly what is approved, under what license, and when it was vetted.

When to use it

Use it when you need a durable, browsable record of vetted models for onboarding, audits, or to stop engineers from re-litigating a model someone already cleared. It complements the PR-time gate by capturing the outcome of every passing check.

How it works

  1. 1A GitHub pull request merged event triggers the run.
  2. 2Hugging Face model IDs added in the merged PR are extracted from the diff.
  3. 3The Hugging Face API confirms each model's license and task tags from its model card.
  4. 4A logic step keeps only models whose license is on the approved allowlist, so nothing non-compliant is recorded.
  5. 5Each approved model is upserted into the Notion registry with its license, task, source PR link, and approval date — updating the row if the model already exists.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect GitHubRepos, issues, pull requests, actions.
  2. 2
    Connect Hugging FaceModels, datasets, spaces — the open-source hub.
  3. 3
    Connect NotionPages, databases, comments.
  4. 4
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  5. 5
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  6. 6
    Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.

Run this workflow in your colony.

14-day trial. No DevOps. No Sales call. Provisioned in under a minute.