Control the document as an ordered set
Source set
A numbered inventory of every page, insert, foldout, and blank side.
Page relationship
Page order and repeated terms connect meaning across page breaks.
Completeness check
Reconcile the final translation against the page inventory before calling it complete.
Inventory before extraction
Inventory first, preserve cross-page context, and reconcile every page before treating the translation as complete.
Create an inventory first
Count the physical pages and name images in reading order before uploading anything. Record blank backs, inserts, foldouts, and attachments. Include page numbers in each capture where they exist.
Photograph pages with the same orientation, distance, and lighting. Retake a weak page immediately rather than discovering later that one paragraph is unreadable.
Preserve cross-page context
Keep repeated headers, section numbers, defined terms, and the last line of the previous page available to the translator. A sentence may continue across a page break, and a pronoun may refer to a heading several pages earlier.
Build a glossary for recurring names and technical terms. Review repeated phrases for consistency without forcing identical wording where the source meaning changes.
Columns and repeated page elements can corrupt sequence, so use the method to fix scrambled OCR reading order before joining pages.
Run completeness checks
Compare the final page list with the initial inventory. Confirm that every heading has content, every footnote marker has a note, and no line was duplicated at an overlap. Spot-check numbers and names on every page.
Contracts, records, certificates, and official submissions may require certified translation or an approved process. A collection of automated page translations is not a substitute for that requirement.
Plan a page-by-page photo translation workflow with stable file order, repeated headers, cross-page sentences, and quality checks. Continue with the documents and layout guide collection for adjacent workflows.