Review model revisions before a deployment change

REAL RECORDS / DATED EVIDENCE

Did the repository metadata change beneath our model watchlist?

Saved at 2026-10-18T13:05:27.991Z from build 1.0.2. This preview makes no live requests.

FROM DATA TO A REVIEW QUEUE

Five practical steps

1. Start with exact public model repository IDs. These two embedding models are examples, not a performance ranking.
2. Record revisionSha alongside the version pinned in your deployment. Fetching metadata does not download or execute model files.
3. Treat licenseDeclared as a publisher label; inspect the repository license and model card for your intended use.
4. Upsert the review row by modelId. Preserve old revision hashes in a separate audit trail rather than overwriting your deployed-version record.
5. Enable comparePrevious with a unique snapshotName for sequential comparisons. Review file-path changes before any download or deployment; this workflow performs neither.

Destination key: modelId. Keep an audit trail and preserve prior rows when a later sample is empty.

TWO WAYS TO START

Run in Apify or download the starter

Use Apify Console

1. Open the Actor, switch to JSON input and paste the example below.
2. Review current pricing, use your own account, and set a $0.05 maximum run charge.
3. Run manually. Review SUMMARY and results before exporting CSV into your review sheet.

Exact sample input

{
  "modelIds": [
    "sentence-transformers/all-MiniLM-L6-v2",
    "BAAI/bge-small-en-v1.5"
  ],
  "maxResults": 2,
  "maxRunSeconds": 120
}

Download a repeatable runner

Save these three files in one folder. The Node.js 22+ starter needs no npm packages. Add your own Apify token through your environment; the downloads contain no credentials.

node run-workflow.mjs --run

Execution is explicit and can incur charges. The starter saves local CSV/JSON, checks run status and coverage, and supports resuming an existing run. Connect your own spreadsheet after reviewing the output. It does not enable a schedule or send notifications.

A SMALL, EXPLAINED START

What this sample would cost at listed event rates

2 results × $0.003 per result, plus $0.00005 Actor Start at the FREE-tier rates observed October 18, 2026. This is a modeled event charge, not a bill or a complete account-cost estimate. Platform/API/storage usage and plan rates may differ.

The runner requests a $0.05 maximum charge. A cap can stop collection early. Check current pricing ↗.

Keep the source’s meaning

• Repository metadata does not establish model quality, security, license suitability or compatibility.
• Artifact counts describe returned paths, not model size or model capability.
• This bounded watchlist does not discover all models or infer repository deletion.

Read the full Actor guide for fields and scope.