Flow — Document lifecycle · OCR path

Summary

A document detected as a scanned PDF (needs_ocr) waits for the user to re-upload a sharper scan or to trigger rescanDocument; the parser then re-runs.

Audience

end-user

Entry

Document marked needs_ocr from the document-lifecycle-ingest branch or document-lifecycle-upgrade-backfill.

Steps

#	Step	Surface	Notes
1	Document is needs_ocr	review-document	Inbox shows "Sharper scan needed" with a Sync button
2a	User re-uploads sharper scan	inbox-uploading	Storage trigger re-runs the paid branch of document-lifecycle-ingest
2b	User taps Sync (rescanDocument)	review-document	Parser re-runs in place
3a	Still scanned	review-document	Stays needs_ocr
3b	Unknown form type	review-confirm-doc	→ needs_classification — see document-lifecycle-classification
3c	Parsed OK	review-document	→ unreviewed — see document-lifecycle-standard-review

Exit

Document moves to unreviewed (or needs_classification); the standard review path can run.

Escapes

• Confirm gate: needs_ocr is NOT confirmable (fields are absent). Confirm button hidden — see the shared rule in document-lifecycle-ingest.
• Repeated still-scanned outcome: user must obtain a sharper scan; no auto-resolution.
• Encrypted/locked PDF surfaces: routes to review-document-locked instead.
• Terminal rejected_format: if rescan reveals an unparseable container, transitions per shared terminal-states table in document-lifecycle-ingest.

Sources

• .context/wiki/flows/* § "Flow 12 — Document lifecycle state machine" · "C. OCR path"