Briefed N4 as batch-convert-scanned-pdf-to-word-with-ocr — raster OCR to one sectioned Word file, not a JPG-batch accept swap or a text-layer PDF doorway.

Date: 2026-09-08 21:25 Summary: Briefed N4 as batch-convert-scanned-pdf-to-word-with-ocr — raster OCR to one sectioned Word file, not a JPG-batch accept swap or a text-layer PDF doorway. Visibility: people

[question] 立项batch convert scanned pdf to word 立项batch convert scanned pdf to word with ocr

[try to solve]

Context

Planner N4 was a deferred scanned-PDF→Word job (ocr pdf to word is a 1k–10k head). The site already has A2 turn-pdf-into-word-document (digital text layer, no OCR) and N4-jpg-batch (photos→one DOCX, no PDF). The user named a batch + OCR task sentence. A second URL turn-a-scanned-pdf-into-word would be a doorway of the same job.

Process

1. Copied the work-tasks template to work-tasks/batch-convert-scanned-pdf-to-word-with-ocr/ (00–03 + notes.md).
2. Locked slug/H1 to the user sentence. Head terms ocr pdf to word / pdf to word ocr absorb as “also called”, not H1. convert scanned pdf to word absorbs on the same page (one PDF still uses the queue).
3. Wrote a control-level IG table vs A2, N4-jpg-batch, N1-batch, images-to-word, and unbuilt N2/N3. Interaction: PDF-only dropzone, Convert all / Download Word, serial pdf.js + Tesseract, skip failed pages, one DOCX (Heading 1 = filename, page breaks, optional page image default on). Caps: 5 files, 20 pages, 12 MB. No auto-loadSample (Tier 2 WASM).
4. Filled 0b coverage + Ads/Planner table + intent review. coverage:gate --phase=0b green. 02=ready, 03=briefs-ready.
5. Updated ocr-scan planner/README, keyword pool (old N4 candidate → absorb; new build row), and tracker.

Root cause / analysis

Searchers who want scanned PDF→Word are not served by extracting a text layer (A2) or by OCR on JPG files. Extra Information Gain must be PDF rasterization, page skip, and one headed Word document. Changing N4-jpg-batch accept to PDF would be scaled-content doorway. Building both turn-a-scanned-pdf-into-word and this batch slug would split one intent.

Solution

Implement later from this brief only: page.style: opts, vendor pdf.js + tesseract + docx, reject images at add-time, FAQ names neighbor H1s. Do not create the old planner slug. Do not attack pdf to word as H1 (A2). Do not occupy week long_gap KPI without SERP.

Notes / boundaries

• No Page.ts / i18n this round.
• CROSS_TOOL FAQ on A2 and N4-jpg-batch waits for implementation.
• Organic SERP titles for ocr pdf to word still need a human look before ship.

[actions]

• Added work-tasks/batch-convert-scanned-pdf-to-word-with-ocr/
• Updated docs/seo/keywords/ocr-scan/*, docs/seo/keyword-daily-pool.tsv, docs/seo/keyword-to-tool-tracker.md
• npm run coverage:gate -- --slug=batch-convert-scanned-pdf-to-word-with-ocr --phase=0b