References
Collect source links
Build a reference list from reports and research documents.
A useful PDF task, made simple
Extract Links from PDF.
Bring the useful links out of your PDF.
Advertisements
Ad placeholderResponsive · up to 1933 × 500Responsive · up to 1933 × 250468 × 60Up to 320 × 100Up to 300 × 50
Extracted data
| Type | Value | Count | Pages |
|---|
Copy and download include all matching rows, across every results page. CSV includes counts and PDF page numbers; TXT contains values only.
Select a page number in Results to check its source. Highlights show approximate positions.
100%
Your extracted list is ready
Your list is ready. Download it, or go back to review and change the format.
Advertisements
Ad placeholder336 × 280468 × 60Responsive · flexible width × 250Up to 300 × 50
Your PDF is processed in this browser.
PDF
How to extract links from pdf
- 01
Choose one PDF
Load a document with selectable text, or try the sample.
- 02
Extract and review
Select Extract. Review values, counts and source pages; filter or sort the list.
- 03
Copy or download
Copy the values, or download CSV with page references or a simple TXT list.
Extract PDF links without opening every page
Collect printed HTTP/HTTPS addresses, www addresses, and clickable web-link targets from one PDF. Keep page references and download a deduplicated list as CSV or TXT.
QUICK VIDEO TUTORIAL
How to extract links and URLs from a PDF
Keep the useful web links without opening them one by one.
English narration
Watch the video
When is this useful?
Resources
Save useful destinations
Extract the destinations behind clickable resource labels.
Review
Keep an audit list
Record where each link was found before reviewing it elsewhere.
Features and controls
Find visible and clickable links
Include a link target even when its label only says “Read more”.
How link extraction works
Printed links beginning with http://, https:// or www. are recognized. PDF link annotations can also reveal HTTP/HTTPS targets behind ordinary text. The PDF does not need to display the URL as its clickable label.
Identical printed targets and annotations on a page are not double-counted. Normalized URL forms are used for duplicate comparison, while the first recognized spelling is displayed. Long, wrapped or damaged URLs may need manual correction. No extracted URL is fetched automatically.
Review values with their source pages
Check counts, jump to a PDF page, and filter the result list.
Review and organize
Each result records the physical page position in the PDF, which can differ from printed page numbers. Select a page chip to open the preview with an approximate highlight.
Remove duplicates starts enabled. Counts combine recognized repetitions of the same value; switch it off to see separate occurrences. Sort alphabetically changes the list order. Filtering applies to copying and downloads as well as the visible rows.
Copy values or download a structured list
Choose CSV for counts and pages, or TXT for one value per line.
What each download contains
CSV contains Type, Value, Count and Pages columns. UTF-8 encoding supports Arabic and other scripts. Values that could be interpreted as spreadsheet formulas are prefixed with an apostrophe for safer importing.
TXT and Copy all contain values only, one per line. All rows matching the filter are included, even when the on-screen table spans several result pages. No new PDF is created and the source PDF is unchanged.
Keep large extractions manageable
Process one PDF in your browser, with explicit limits and partial-result notices.
Files, limits and scanned pages
Load one PDF up to 50 MiB and 500 pages. Extraction checks up to 250,000 characters and 2,000 matches per page, 10 million characters and 20,000 matches overall, and 2,000 annotations per page. A 90-second extraction budget limits lengthy runs. Very complex documents may work better when split into smaller files.
When a limit is reached or a page cannot be fully read, the results show a notice and limited or failed runs export with partial in the filename. Pages without selectable text are counted separately. Use PDF OCR for image-only scans, then return to extract the recognized text.
PDF content is processed locally in this browser. Uploaded documents are not sent to an extraction server. Links, email addresses and phone numbers found inside the PDF are never contacted automatically.
Supported files & processing
- Input
- Output
- CSV / TXT
Open one PDF up to 50 MB and 750 pages. Additional output, memory and processing limits can apply to complex documents.
Processing takes place on your device without uploading the document. Download your result before closing the page; keep the original separately.
Advertisements
Ad placeholderResponsive · up to 1933 × 500Responsive · up to 1933 × 250468 × 60Up to 320 × 100Up to 300 × 50
FAQ
A few useful answers
Know what to expect before you start.
On your deviceWhat can this tool recognize?
Collect printed HTTP/HTTPS addresses, www addresses, and clickable web-link targets from one PDF. Keep page references and download a deduplicated list as CSV or TXT. The tool does not check whether links still work. Internal page destinations, JavaScript actions, mailto and other non-web schemes are not exported as web links.
Can I extract data from scanned PDFs?
Only selectable text and recognized link targets are read. Image-only pages need PDF OCR first. Run OCR, download the searchable PDF, then return to this extractor.
How are duplicates and counts handled?
Remove duplicates combines recognized equivalent values. Count shows detected occurrences and Pages lists physical PDF page positions. Disable the option to keep occurrences separate. A printed target and matching clickable link on the same page are not double-counted.
What is the difference between CSV and TXT?
CSV includes type, value, occurrence count and page references. TXT and Copy all provide values only, one per line. All rows matching the current filter are included, not only the visible table page.
Is the PDF uploaded for extraction?
No. PDF parsing, matching and exports run in this browser. The original PDF is unchanged. No extracted link, email or phone number is contacted automatically.
What are the limits?
One PDF up to 50 MiB and 500 pages. Extraction is limited to 250,000 text characters and 2,000 matches per page, 10 million characters and 20,000 matches overall, and 2,000 annotations per page. Results carry a notice if limits or unreadable pages make them partial.
