Developers
Feed a XML processing workflow
Export page content as XML for a script or application that understands the documented output structure.
A useful PDF task, made simple
PDF to XML.
Export PDF text, images and drawing geometry as XML.
Advertisements
Ad placeholderResponsive · automatic sizeResponsive · automatic size468 × 60Up to 320 × 100Up to 300 × 50
Use the page controls to inspect your document.
100%
Your file is ready.
Download your result, or return to the settings to make another version.
Advertisements
Ad placeholderResponsive · automatic size
Sent for temporary processing when you start.
PDF
How to pdf to xml
- 01
Choose your document
Upload one PDF from your device.
- 02
Set up the result
Review the available settings, then start processing.
- 03
Download your file
Save the result and check it before sharing.
PDF to XML
Export PDF text, images and drawing geometry as XML. Select pages and download a structured description, with image assets packaged alongside it when needed.
QUICK VIDEO TUTORIAL
How to convert PDF to XML with layout data
Get structured page data for your own workflow.
English narration
Watch the video
Practical ways to use this tool
Researchers
Inspect text positions
Use content coordinates to study page layout; they refer to the unrotated PDF page rather than browser pixels.
Technical review
Separate extraction from interpretation
Review extracted objects before building document-specific parsing rules. Scans need OCR for recognized text.
Features and controls
Export text and drawing geometry
Retrieve a structured description of page content rather than a plain transcript.
Schema, coordinates and page rotation
The versioned structure includes pages, text blocks and spans, fonts, bounding boxes and vector drawing paths. Coordinates are in PDF points from the unrotated page’s top-left origin; rotation is recorded separately. This can support indexing or document analysis, but the converter does not infer your own business schema or semantic field names.
Keep image assets beside the data
Receive a direct data file or a ZIP when extracted images are included.
Data file and referenced images
When a page contains image data, the archive contains document.json or document.xml plus referenced image files and any supported masks. Keep those paths together when loading the data elsewhere. An image-only scan remains image data; this export does not perform OCR to invent missing text spans.
Choose the pages you need
Process the whole document or enter specific page numbers and ranges.
Page ranges and numbering
Use PDF positions, starting at 1, rather than numbers printed on the page. A list such as 1, 3-5 includes four pages. A blank range uses every page. Check the preview when the document includes a cover or differently numbered chapters. Only the chosen pages are included in this conversion.
Use OCR for image-only pages
Text extraction reads a text layer; it does not recognize letters in images.
Scans, empty pages and OCR
Extraction uses an existing text layer and does not recognize letters in page images. If a page has no readable text, use PDF OCR separately. A blank page can also produce no text. For mixed PDFs, inspect empty or unreadable pages individually. OCR processes one selected page at a time.
Supported files & processing
- Input
- Output
- XML, ZIP
Open one PDF up to 100 MB and 750 source pages. Protected files need an unlocked copy unless this tool explicitly offers password entry. Large or complex pages can require more memory or reach processing limits even when the file is within these limits.
Processing uses our servers. Source files, temporary files and results are deleted from our servers within 30 minutes after processing finishes. We do not share document contents with third parties or use them to train AI models.
Advertisements
Ad placeholderResponsive · automatic sizeResponsive · automatic size468 × 60Up to 320 × 100Up to 300 × 50
FAQ
A few useful answers
Know what to expect before you start.
Temporary server processingDoes PDF to XML recognize invoice fields or table meanings?
It exports page content and geometry, not automatically named business fields. Your application must interpret that structure. For table cells, use Extract Tables from PDF; for scans, use PDF OCR.
What information does the structured export contain?
The versioned structure includes pages, text blocks and spans, fonts, bounding boxes and vector drawing paths. Coordinates are in PDF points from the unrotated page’s top-left origin; rotation is recorded separately. This can support indexing or document analysis, but the converter does not infer your own business schema or semantic field names.
Why does the download sometimes contain a ZIP?
When a page contains image data, the archive contains document.json or document.xml plus referenced image files and any supported masks. Keep those paths together when loading the data elsewhere. An image-only scan remains image data; this export does not perform OCR to invent missing text spans.
Will this extract words from an image-only scan?
Extraction uses an existing text layer and does not recognize letters in page images. If a page has no readable text, use PDF OCR separately. A blank page can also produce no text. For mixed PDFs, inspect empty or unreadable pages individually. OCR processes one selected page at a time.
What are the file and page limits?
Open one PDF up to 100 MB and 750 source pages. Protected files need an unlocked copy unless this tool explicitly offers password entry. Large or complex pages can require more memory or reach processing limits even when the file is within these limits.
Where is my document processed?
This tool uses our servers for processing. Your original file is not overwritten; the result is a separate download. Preview and review the result where available, and download the copy you need before starting over.
