ID documents

Convert an ID card scan to Word without flattening the layout

A card is not a letter. Photo, headings, and field rows sit in a tight grid. Ordinary PDF to Word tools treat that as a column of text and the card falls apart. Rebuild Page has a dedicated ID path: reconstruct the card as HTML, review it next to the scan, then export Word from that HTML.

What this is for

Photo ID cards, licenses, and similar card-sized scans where you need an editable Word file that still looks like the original. Typical inputs are a phone photo, a flatbed scan, or a one-page PDF of the card.

This is a conversion tool for documents you are allowed to process. It is not a way to forge, bypass KYC, or extract someone else's identity data.

Why generic converters fail on IDs

Most converters optimize for reports and letters: they linearize text and drop geometry. On an ID that means the portrait floats, labels detach from values, and bilingual lines merge. The ID path keeps those regions as a layout, then lets you correct OCR in the HTML studio before Word is built.

How to convert an ID

1. Sign in and upload the PDF or JPEG/PNG of the card.
2. Wait for the job to finish. Auto-detect should pick the ID type.
3. Compare reconstructed HTML to the original. Fix names, numbers, or labels if needed.
4. Download Word from that HTML.

The same steps are documented for products on the Partner API page — use document_type: id. For the full HTML-then-Word pipeline, see how it works.

Credits and retention

ID pages use a higher per-page hold than a simple letter, then settle to measured usage. Files are removed after 14 days unless you keep them. Pricing and privacy have the details. We do not train models on your uploads.

Questions

Can I upload a photo of an ID card, not a PDF?

Yes. JPEG and PNG scans work. Upload the card image the same way you would a PDF page.

Will the photo and field labels stay in place?

That is the point of the ID path. Rebuild Page reconstructs the card as HTML so portrait, labels, and values keep their layout, then builds Word from that HTML.

Do you keep ID scans forever?

No. Uploads and outputs are deleted after 14 days unless you mark the job Keep. You can delete immediately from the job page. Documents are not used to train models.

Is this the same as document_type=id on the API?

Yes. The studio auto-detects ID-style pages. Partners can set document_type to id on POST /v1/jobs and download DOCX from the reconstructed HTML.