Skip to main content

End-to-End Workflow

Putting the pieces together, a typical metadata-QA cycle for a catalogue of records looks like this:

  1. Configure a client once, choosing a provider (Quick Start).
  2. Submit each record and collect its results. For many records, submit them and wait in turn, or submit a batch and wait on each Job handle.
  3. Assemble a workbook with one row per record and the issue array as a JSON string (Review Board).
  4. Triage in the review board — a human opens the workbook, works through projects highest-severity-first, and confirms which suggestions to apply (Review Board).
  5. Apply accepted fixes back to your metadata system as a separate step, using the key paths in current_metadata / suggested_metadata to target each field.

A minimal batch driver that submits everything first, then gathers results:

jobs = {name: client.submit(md) for name, md in catalogue.items()}

results = {}
for name, job in jobs.items():
    try:
        results[name] = job.wait_sync(timeout=300)
    except RuntimeError as exc:
        print(f"{name}: {exc}")
        results[name] = []
Throughput tip

Submitting many jobs at once issues many concurrent LLM calls, which can hit provider rate limits. If you see rate-limit errors surface as failed jobs, submit in smaller waves, or run sequentially (submit, wait, repeat) to keep concurrency low.