Per-page OCR decisions
Eligible zero-text image pages can receive the current English OCR lane. Existing usable text passes through, and searchable text is checked in the delivered PDF.
Loading...
Loading...
Documents is one of four equal pillars alongside LMS, web, and media. Aelira scans PDFs and may apply eligible changes when the file exposes a safe target; it does not turn every finding into an automatic fix.
The PDF hardening described here shipped in the immutable Aelira Core v0.9.6 release on 25 August 2026. It is available to self-host from that release; this page does not claim the behavior is live in Aelira's hosted service.
Capability depends on the page, issue metadata, structure tree, installed dependencies, and verification result. These are bounded implementation paths, not outcome guarantees.
Eligible zero-text image pages can receive the current English OCR lane. Existing usable text passes through, and searchable text is checked in the delivered PDF.
Aelira may update metadata, bookmarks, and eligible heading, list, table, structure, and figure alt-text targets when the PDF exposes enough unambiguous context.
Output candidates are validated before managed publication. Re-scan and integrity evidence support review; they do not certify PDF/UA or WCAG conformance.
Step 1
Inspect text, OCR state, metadata, tags, reading-order signals, images, tables, links, forms, and bookmarks.
Step 2
Apply only eligible fixes. Unsupported, ambiguous, refused, or unverifiable work remains manual or failed.
Step 3
Proofread OCR and alt text, inspect structure and reading order, compare pages, and exercise interactive behavior.
Step 4
Use Aelira evidence plus the external validators and assistive technology required by your institution.
Agree the corpus, supported cases, human-review workflow, acceptance evidence, and limits in writing.