Translate PDF documents with Hyperlocalise - Hyperlocalise

Hyperlocalise reads the native text layer of a PDF, creates one segment per text block, and writes approved translations back into a copy of the source file. Supported extension: .pdf. Scanned PDFs without selectable text fail ingest.

What is translated

Text drawn on each page as PDF text objects, grouped into blocks by layout.
Keys look like page-1#1 (page 1, block 1). Context reads Page 1, block 1.

Text inside form XObjects, annotations, and form fields is not extracted.

Target PDF

The CLI writes a target application/pdf from ready segment translations. Blocks that cannot fit after wrapping and font scaling are still written with the best-effort layout, then listed as overflow keys.

Configure your files

Use one PDF per locale with a path template:

files:
  - source: content/handbook.pdf
    targets:
      - locales: [fr-FR, de-DE]