Common editorial and ops failures for topline extraction.
Extraction looks incomplete
The PDF may be scanned imagery. Claude can OCR images, but low-resolution scans lower confidence. If confidence is below 0.70, ask engineering to force a re-extraction.
Status stuck on pending
Jobs run on Action Scheduler. Wait 2–5 minutes and refresh. After 10 minutes with no change, the queue may need ops attention.
Extraction URL returns 404
Confirm the pdf_extraction post is publish and the parent report slug has not changed. Public paths are /{parent-slug}/extraction and /{parent-slug}/text.
Need a fresh re-run
Editors cannot force overwrite from the UI. Ops run:
wp prc pdf-extraction process --post_id=<ID> --forceRelated
- User guide
- Architecture
- Plugin README — full WP-CLI list