AI / upload mode roadmap

OCR PDF Roadmap

OCR is a planned advanced feature. Lightweight local OCR can work for simple files; high-accuracy OCR may require explicit upload mode.

Review OCR roadmap Full toolkit

how it works

Designed for the shortest private path.

Use local tools for current PDF work.
OCR Lite will be evaluated for browser processing.
OCR Pro will require explicit upload consent if shipped.

advanced feature boundary

OCR PDF is not shipped in the local MVP yet.

This feature needs a higher quality bar than the current browser-only pipeline. If it later requires upload or AI processing, VaultPDF will separate it from local tools and ask for explicit consent before any file leaves the browser.

Use local tools now

quality gate

OCR quality gate

Searchable text layer with usable reading order on common scans and clear disclosure for handwriting or low-quality scans.

hybrid candidate

Must pass before launch

Measure character error rate on invoices, contracts, receipts and academic scans.
Preserve page count and visual alignment after adding text layer.
Disclose local OCR Lite versus upload-mode OCR Pro before processing.
Reject files that exceed local memory limits instead of freezing the page.

Currently blocked by

No representative OCR fixture set yet.
No accuracy benchmark report yet.

seo content matrix

Advanced AI and conversion roadmap

Use explicit upload labels, retention policy and accuracy criteria before launch.

Audience: power users who need OCR, Office output, summary and Q&A
Primary intent: understand when upload-mode processing is necessary

ocr pdf roadmappdf to word qualitysummarize pdf with aichat with pdf citations

faq

Practical boundaries before you process a file.

Is OCR available in the MVP?

No. It is intentionally separated because OCR quality and privacy boundaries need careful handling.

Will OCR upload files?

Only OCR Pro may require upload, and it must be labeled before the user starts.