A designer hands over a polished PDF, the translation deadline is fixed, and the InDesign file is nowhere to be found. The PDF may look complete, but that doesn’t mean its text, tables, and layout can be cleanly edited or translated as if the original source were still available.
The first decision isn’t which conversion button to click. It’s what the client needs at the end: translated text, a rebuilt InDesign document, or a proofed PDF with an editable IDML handoff. Those are different deliverables, with different risks and amounts of layout work.
What the missing source file changes¶
A PDF is a format for preserving how a page looks, not a guarantee that the page’s content remains organized like an InDesign document. The pypdf extraction guide explains that PDFs may lack a semantic layer identifying paragraphs, tables, headers, or footers.
A PDF page can show a table while storing its contents as separate pieces of text at specific coordinates. A text extractor may see those pieces, but it can’t automatically know which values belong in the same row or which line is a heading. Columns, footnotes, captions, and page numbers can raise similar questions.
The pypdf documentation puts it plainly: “PDF files don’t contain a semantic layer.”
That distinction matters because translation needs more than visible words. A translator or production team also needs to understand what each fragment does, which elements belong together, and how the translated text should flow through the layout.
A selected-and-copied paragraph can still be incomplete or scrambled. Extraction may make decisions about line breaks, repeated headers, ligatures, whitespace, and reading order. Even when the output looks close to the page, a plain text file isn’t automatically a safe substitute for an editable source.
A question from an InDesign user shows how easily the task can be misunderstood. The user wrote, “I have Indesign 20.5 and read, that you can open pdf documents in this version now and edit them.” The Adobe Community question, posted October 22, 2025, describes that user’s experience, not a general conversion guarantee.
The practical distinction is simple: placing a PDF into an InDesign document gives you a page reference, not the original editable text and layout structure. A placed page can help with visual reconstruction, but it doesn’t bring back the original paragraph styles, text frames, table relationships, or linked assets.
A second Adobe Community post describes the familiar handoff gap: a designer supplied a book PDF but not the original InDesign file. The question posted November 11, 2021 is an example of the problem, not proof that one conversion method will work for every book.
Treat the PDF as a visual reference until a sample-page check shows what can actually be extracted. If the client later finds an INDD, IDML, earlier translation, or editable text file, the project may become simpler. Ask before estimating reconstruction work.
Start with the deliverable and the available files¶
The same source PDF can lead to several different projects. A translation-only handoff may not need an InDesign rebuild at all; a print-ready localized edition usually does. Agreeing on the result first prevents the team from spending time rebuilding pages the client never intended to edit.
Ask the client for the original INDD or IDML file, packaged fonts, linked graphics, editable text, and any earlier language version. One missing file can matter more than another: an earlier language edition may reveal how the designer handled recurring labels, tables, or page furniture, while packaged graphics can save time recreating illustrations.
Use this checklist in the project brief:
| What to ask for | Why it matters | What to agree if it’s missing |
|---|---|---|
| INDD or IDML source | May preserve editable text, styles, frames, and page structure | Whether to reconstruct the source layout or build only the target edition |
| Packaged fonts | Lets the designer use the intended typefaces | Whether to use substitutes and who approves the visual change |
| Linked graphics | May provide editable or higher-quality original assets | Whether to extract, replace, or leave graphics unchanged |
| Editable text or earlier language version | Helps recover wording and repeated terminology | Whether PDF extraction or OCR will be the text basis |
| Reference PDF | Shows the page appearance to match | Which PDF version is final if the client has multiple copies |
| Delivery specification | Defines what the client can edit after handoff | Whether delivery includes translated text, IDML, PDF, or a package |
The project brief should also record the source and target languages, intended audience, page size, output purpose, and who approves terminology and visual changes. Specify whether the client expects the target edition to follow the original line by line or whether the designer may adjust spacing and page composition to fit natural target-language text.
Make the scope concrete. “Translate the PDF” could mean any of the following:
- Provide translated text in a separate document.
- Translate text and return it in a document that roughly follows the source layout.
- Reconstruct the source-language layout in InDesign, then translate that editable document.
- Create a target-language layout and deliver a proofed PDF.
- Deliver an editable IDML file, and specify whether fonts and linked graphics are included separately.
Those outcomes shouldn’t be treated as interchangeable. An IDML handoff can preserve useful InDesign structure, but a PDF alone doesn’t supply the original source package. For a broader discussion of what belongs in a production handoff, see client-ready documents vs translated text.
Run a feasibility check on representative pages before quoting a full reconstruction. Include a plain text page, a page with columns, a table, and a scan if the document has them. Check whether text is selectable, whether extraction follows the visible reading order, and whether footnotes, captions, and repeated headers stay attached to the right content.
The pypdf guide describes why extraction can involve judgment about line breaks, tables, and reading order. A sample-page review is a practical way to discover those issues before they spread across a whole document.
| Sample-page check | What to inspect | What a problem could mean |
|---|---|---|
| Selectable text | Can you select a sentence as text? | A scan may need OCR before translation |
| Reading order | Does extracted text follow the page from top to bottom and across columns? | Manual correction or structured reconstruction may be needed |
| Tables and callouts | Do labels and values remain together? | Extraction may need repair before translation |
| Repeated page elements | Are headers, footers, and page numbers distinct from body text? | Page furniture may be duplicated or translated incorrectly |
| Fonts and characters | Do special characters and names survive extraction? | Check the visible reference and agree on source text |
| Visual complexity | Are there sidebars, captions, or overlapping elements? | Layout work may require manual recreation |
Don’t promise a faithful editable conversion based only on selectable text. Text selection tells you that characters can be extracted; it doesn’t tell you whether paragraphs, tables, or page relationships will survive in a useful form.
Choose between two production workflows¶
With no native source file, the team usually chooses between translating extracted text before rebuilding the target layout and rebuilding the source layout first. Neither path suits every PDF. Layout complexity, expected revisions, language count, and the need for CAT-ready IDML all affect the choice.
Workflow A: extract, translate, and build the target layout¶
This route separates linguistic work from page reconstruction. The team extracts text from the PDF, corrects extraction errors against the visible page, translates the verified text, then builds the target layout in InDesign.
The route can suit a simple brochure or a short document with mostly continuous text and few complex tables. The translator can work on text in a clear sequence while the designer prepares the target page structure. The team still needs to map text back to its original locations and account for page furniture and visual elements.
A safer sequence looks like this:
- Extract selectable text or run OCR on scanned pages.
- Compare the result with the visible PDF and correct errors.
- Mark headings, captions, footnotes, tables, labels, and repeated elements.
- Translate the corrected text with agreed terminology.
- Rebuild the target layout and adjust it for the translated wording.
- Review the final pages for language, order, overflow, and visual consistency.
The main advantage is that translation can begin without first recreating every source-language page. The main risk is losing relationships between extracted text fragments and their positions, especially in tables, multi-column pages, and designs where text wraps around images.
A translator who receives a text dump also needs context. A short label like “Open” could be a button, a status, or an instruction. Capture screenshots or page references for ambiguous text, and label content with page and element details rather than sending one unmarked stream.
Workflow B: reconstruct the source layout, then translate¶
This route starts by recreating the source-language pages as editable InDesign content. The team rebuilds text frames, tables, styles, and image placement, checks that the reconstruction matches the reference, and then translates the editable layout.
Rebuilding first can be a better fit when text relationships matter, the layout is dense, the client expects later revisions, or a CAT-ready IDML handoff is part of the brief. A CAT tool is translation software that manages text segments and terminology. An editable handoff can make repeated updates easier than returning to a PDF and extracting everything again.
The initial reconstruction takes work, and the reconstructed file still isn’t the original native source. Fonts or linked graphics may be missing. The designer may need to substitute fonts, recreate elements, or ask the client to approve changes to assets that can’t be recovered.
For teams comparing layout-first production with a simpler document workflow, InDesign vs Word for multilingual layout explains when each format makes sense. A related IDML round-trip workflow covers handling an editable InDesign handoff through translation and back into layout.
| Decision factor | Extract, translate, then rebuild | Rebuild source layout, then translate |
|---|---|---|
| Simple pages with continuous text | Often a practical fit | May add reconstruction work before translation |
| Dense tables or complex page structure | Higher risk of losing relationships in extracted text | Lets the team define structure before translation |
| Need for repeated revisions | Each change may require careful remapping | Editable text and styles can support later updates |
| Need for a CAT-ready file | Depends on how text is prepared and handed off | Better aligned with an editable IDML handoff |
| Missing fonts or graphics | Affects the target rebuild | Affects source reconstruction and may require substitutions |
| Priority is a translated text file | May avoid a full source-layout rebuild | Could be more layout work than needed |
A third path is to deliver translation without reconstructing InDesign at all. That option can work when the client only needs reviewed text, but it isn’t a finished localized layout. State that distinction in the quote so nobody expects an editable, print-ready file from a text-only delivery.
The PDF standard, ISO 32000-1:2008, describes the PDF format itself. For the practical extraction question, the pypdf documentation is more direct: the visible page doesn’t necessarily encode the structure a layout editor needs.
Handle scans and OCR before translation¶
A scanned PDF is a page image, not a page of selectable text. OCR, or optical character recognition, attempts to read text from that image and create a text layer. OCR can give the team something to edit, but it doesn’t recover the original InDesign document or prove that every character has been read correctly.
The pypdf documentation notes that image-only or scanned pages may yield little or no extracted text, and that pypdf isn’t OCR software. A scan therefore needs a separate recognition step before the text can be checked and translated.
Check the OCR result against the page before translation. Focus on:
- Names, numbers, dates, and reference codes.
- Tables, including column order and labels.
- Headers, footers, captions, and footnotes.
- The order in which text should be read.
- Stamps, signatures, seals, and low-contrast areas.
- Characters that can be confused in the scan.
OCR text is a transcription layer, not evidence that the original layout has been recovered. A scan can contain clear words but still leave the team unsure where a label belongs, whether a number sits in a table cell, or whether a mark is a stamp or a graphic.
For pages that combine OCR with substantial cleanup, ChatsControl is one possible document-translation step. Uploading a scanned PDF runs text recognition and produces a formatted Word file that follows the page layout; the result can be reviewed before anyone uses it as a basis for an InDesign rebuild. The service bills scans by recognized characters at the same rate as text documents, with at least 1,000 characters per page. Very poor scans and handwriting remain unreliable, so a human check against the original is still needed.
That workflow may help when the immediate problem is extracting and reviewing content, but a formatted Word file isn’t a reconstructed InDesign source. The designer still needs to build or adjust the target layout and confirm the final pages visually.
If the PDF is an image-only file, use the more detailed guide to translating an image PDF with OCR. For projects that involve CAT-tool preparation, file preparation for translation can help the team plan what to include in the translator handoff.
Rebuild the target layout and check the translation¶
Translated text won’t always occupy the same space as the source. Longer or shorter wording can change line breaks, push text beyond a frame, alter the balance of a page, or move content to another page. Don’t treat the source PDF’s coordinates as fixed boxes that translated strings can simply be dropped into.
The designer should preserve the content hierarchy while making layout changes that keep the target text readable. Check line breaks, hyphenation, spacing, page count, and text overflow. If a paragraph moves, check whether its caption, image, footnote, or table still appears with it.
A useful production review has separate language and visual passes:
- Completeness: Confirm that every source element has a target equivalent or a documented reason for omission.
- Reading order: Compare the intended sequence with the final pages, especially in columns and sidebars.
- Text placement: Check that headings, captions, labels, and footnotes sit beside the correct elements.
- Tables and figures: Verify values, labels, numbering, and references against the approved translation.
- Overflow and pagination: Inspect every text frame and the page sequence after layout changes.
- Links and print boundaries: Check links and confirm that content remains within the required page boundaries.
- Deliverable files: Open the agreed PDF and editable file, and state which parts were reconstructed or substituted.
A bilingual review can help a linguist compare source text and translation without switching between separate documents. In ChatsControl, the source and translation appear side by side, with translated text highlighted and the original available on hover. That supports review of the text, but it doesn’t replace the designer’s visual proof of the final InDesign pages.
The pypdf extraction guide notes that text can be positioned in fragments and that extraction methods may offer layout-oriented output.
Layout-oriented output can help a team understand where text appeared, but it still doesn’t make the extracted file a reliable editable layout source. The designer needs to verify the relationships visually and decide what must be reconstructed.
A good proof compares the target PDF against both the approved translation and the source reference. Checking only the translated text can miss a missing graphic or a misplaced caption; checking only the visual match can miss a changed number or omitted footnote.
Quote and hand off the work without hidden assumptions¶
The price and schedule depend on what the team has to reconstruct and what the client expects to receive. A quote based on page count alone may miss the work hidden in scans, tables, columns, graphics, and repeated revisions. Scope the production stages separately so the client can see what changes the estimate.
| Quote item | Questions to settle before pricing |
|---|---|
| File inspection | Which pages need a sample extraction or OCR check? |
| Text preparation | Who corrects extracted text and checks it against the PDF? |
| Translation | Who approves terminology, names, and ambiguous labels? |
| Source reconstruction | Does the project need an editable source-language layout? |
| Target-language DTP | Who adjusts frames, tables, page breaks, and fonts? |
| Review | Are linguistic review and visual proof both included? |
| Handoff | Is the final delivery a PDF, IDML, translated text, or a package? |
| Revisions | How will client wording changes after layout be handled? |
Separate translation from DTP in the scope. Translation covers the wording; desktop publishing, or DTP, covers arranging and proofing the content in the layout. A client may need both, but the PDF alone doesn’t make the DTP work disappear.
Set a change-control rule for revisions. If the client changes approved wording after the layout proof, the team may need to adjust text frames and recheck pagination. Define who approves the final text and how many review rounds are in scope rather than assuming that each revision is a simple string replacement.
Before delivery, state what was reconstructed and what was substituted. Identify missing fonts, replaced graphics, unverified OCR, or content that couldn’t be read from the reference. A transparent handoff gives the client a usable file and a clear record of its limits.
A final package might include the proofed PDF, editable IDML, translation text, and a note about fonts or linked graphics. Don’t label a PDF as an editable InDesign source, and don’t promise that a reconstructed file contains the original styles or assets if the source package was never supplied.
Common PDF-to-InDesign pitfalls¶
The most expensive problems often start with a small assumption: that selectable text is structured text, that OCR is proofread text, or that a placed PDF is an editable InDesign page. Each assumption can create extra correction work once translation or layout has started.
Treating a placed PDF as a converted file. Placing a PDF provides a visual reference. It doesn’t restore text frames, paragraph styles, tables, and linked graphics from the original document. The Adobe Community question about opening a PDF in InDesign illustrates why users can expect a conversion step that the available workflow doesn’t actually provide (Adobe Community, October 22, 2025).
Translating an unverified text dump. A line break or reading-order error can change the meaning of a table or a column. Compare extracted text with the visible PDF before translation, and record page references for content that could be read in more than one way. The extraction decisions documented in the pypdf guide explain why text-only output needs review.
Assuming a scan has no translation cost beyond its visible words. OCR may misread names, numbers, and small text. Check recognition first, especially in tables, footnotes, and stamped areas. An OCR result that looks plausible can still contain a subtle error.
Replacing text in fixed boxes. Translation may need more or less space than the source. A quick replacement can create overset text, poor line breaks, or awkward page turns. The layout needs a target-language proof, not just a search for untranslated text.
Leaving the handoff undefined. One person may expect a translated PDF while another expects editable IDML with fonts and graphics. Record the exact file types, approval steps, and known substitutions before production begins.
The Adobe Community post about a missing InDesign source is a real example of the handoff problem, but it doesn’t establish a universal remedy. The useful lesson is to ask for the source package and define the fallback before promising an editable result.
As the pypdf documentation explains, “The PDF format is meant for printing. It is not designed to be read by machines.”
That quote doesn’t mean a PDF is impossible to translate. It means the production team should distinguish between a page that looks right and a file whose content structure is ready for editing.
FAQ¶
Can InDesign convert a PDF into an editable InDesign document?¶
A PDF can be placed as a visual reference, but placing it doesn’t recreate the original InDesign structure. Editable text, tables, and styles may need manual reconstruction.
Should I translate a PDF before or after rebuilding it in InDesign?¶
Translate extracted text first when the layout is simple and text accuracy can be checked separately. Rebuild the source layout first when structure, repeated revisions, or an editable IDML handoff matters.
How do I translate a scanned PDF for InDesign?¶
Run OCR, then compare names, numbers, tables, footnotes, and reading order with the visible page. Treat OCR as a text transcription, not as recovery of the original InDesign file.
Can I quote a PDF translation before seeing the whole file?¶
You can scope a project after reviewing representative pages, but a sample-page check is needed to assess selectable text, scans, reading order, tables, columns, and layout reconstruction.
What files should the client provide for a translated InDesign handoff?¶
Ask for INDD or IDML, packaged fonts, linked graphics, editable text, and any earlier language version. If none exist, agree whether the deliverable is translated text, a rebuilt layout, or a finished PDF and IDML package.
Will translated text fit the original PDF layout?¶
Not necessarily. Translation can change line breaks and page count, so the target layout needs checks for overset text, hyphenation, spacing, and page boundaries.
Does a text-extraction file count as an editable InDesign source?¶
No. Extracted text may lose reading order, table relationships, and other structure. It can support translation, but it doesn’t replace a reconstructed InDesign layout.