PDF Compare
Compare the text of two PDFs in the browser and see exactly what changed.
Jobs Compare
Operated by Logrun Lab
Opens the vendor's own site.
Jobs covered
Open a task to see exactly what the vendor documentation supports.
Compare
Browse all compare toolscomparing the text of two PDF files in a browser with PDF Compare, where the current official page at https://logrunlab.com/en/pdf-compare says PDF Compare extracts and normalizes the text of two PDFs in the browser and shows mechanical Added, Removed and Changed results with page references, defines Changed as a paired text block existing on both sides with differing wording, Added as text appearing on the New side without a paired block on the Original side, and Removed as text appearing on the Original side without a paired block on the New side, explains that Original and New are positions chosen for the comparison so that Added and Removed are relative to them and that the tool does not determine which file is historically newer or who made an edit, warns that page references do not guarantee visual alignment or complete layout comparison and that the labels do not indicate importance risk or authorship, says that no meaningful text changes found does not prove the complete PDF files are identical and that the comparison does not cover image-only or scanned content without a usable text layer, images and graphics, visual formatting or layout only changes, metadata or other non-text structure, requires an existing usable text layer and does not run OCR while a mixed PDF can be accepted if some pages contain extractable text and image-only pages are not compared, normalizes supported extraction noise such as whitespace and some line wrap or page break effects, limits each PDF to 20 MB and 200 pages, states that password protected or encrypted PDFs are not supported in the current version, and says that PDF files, extracted text and comparison results stay in the current browser session and are not stored by Logrun Lab as document content, with TXT and MD named as formats that may be added in future versions and the linked contract version comparison workflow at https://logrunlab.com/en/compare-contract-versions adding that Logrun Lab compares extracted text and does not read text in images or scans, compare page appearance or interpret legal meaning
Where your files are processed
Files stayed on your device when we tested
We supplied a test PDF on 21 October 2026 and no file left the device. Our check confirmed the tool actually ran: A comparison result was reported, so the tool processed both documents.
It is a browser tool: the official copy supports text comparison in the browser with page references and a 20 MB and 200 page limit per PDF. What the record can say about the network is what a recorded test observed, not that nothing left the device, because the absence of marker text alone is not proof against encoded or transformed transmission, and this audit did not reproduce the test.
What it does
PDF Compare takes two PDFs and reports what changed between their text: Added, Removed and Changed, with page references. It reads both files in the browser and normalises their text layers before comparing, so the output is a mechanical list of differences rather than a visual overlay.
It works on text based PDFs only, up to 20 MB and 200 pages each, and it says so plainly when a file does not qualify. It is the first record filed under the compare job, which was added to this index on 2026-09-29.
Best for
Anyone checking a revised document against the version it replaced, where both are text based PDFs: a contract, a specification, a policy, a set of proofs. It is at its best on long documents where reading both and finding the changes by eye is the slow way.
Worth knowing
It compares text, not pages, and those are different jobs. A scanned document has no text layer and will be refused, and a mixed PDF only has its text bearing pages compared, so a change that exists only as an image or a moved signature is invisible to it.
It is not a visual or pixel level comparison and it does not run OCR. Password protected and encrypted PDFs are not supported.
A comparison also only tells you what the text says changed: rewording a clause so that it means something different but reads similarly is reported as Changed without comment, which is the human half of the job. Text comparison is distinct from visual comparison and from OCR.
The recorded network test is stronger evidence than a vendor promise, but its original trace, fixtures, browser and build and date need preserving before its exact request counts are carried forward. It is not the only product here with a comparison feature: editor records can carry comparison capability without a Compare category, so this is one place to compare documents rather than the exclusive one.
Details, and where each comes from
Vendor statements, our own reading of their pages, and the things nobody has established yet, each answered separately. Where we have not confirmed something it says so rather than guessing, and a row nobody has established is greyed out.
- Versions covered
- Web
- Website
- logrunlab.com
- Requires an account
- No
- Watermark on the free tier
- Not verified
- Maximum file size
- 20 MB
- Batch processing
- No
- Data residency
- Not verified
- Open source
- No
- Self-hostable
- No
- Licence
- Not verified
- Redaction
- No redaction feature
- Programmatic access
- None
- Last availability check
- Responded in 191 ms 21 October 2026
Checked automatically by our own runner, not reported by the vendor. 100% of our last 27 checks got a response.
What the redaction verdict rests on
From the vendor's own documentation, read 29 September 2026. Not tested by us.
What that does not establish
There is no redaction feature of any kind. This is a comparison tool: it reports differences between two documents and does not mark or alter either.
Recorded as none rather than unverified because the absence is stated by the vendor and visible in the tool, which offers only Compare and Reset.
What the vendor says
Claims are labelled rather than repeated as fact. Where we have checked something ourselves, it says so and shows when. Checked means we read the vendor's own page and it says what we recorded - not that we tested the behaviour, unless the note says we did.
-
“PDF Compare extracts and normalizes their text in your browser, then shows mechanical Added, Removed and Changed results with page references.”
Checked by us - the source says this Read 29 September 2026
Confirmed, and tested rather than read. The comparison reported "1 Changed 0 Added 0 Removed" and then "CHANGED Page 1 to Page 1", printing the previous marker and the updated one.
That is precisely the difference the two files were built to contain, reported in the right place, so the tool genuinely read both documents and compared their text. It did not merely accept the files and render something.
Standing: Not re-read against the live source
Check it yourself: the source, the SHA-256, and how we checked
- From
- the raw HTTP response
- SHA-256
- 0947092a798846540e3fe59613b2c9a9c38df5484d583262a28ce5366e5b44b3
How we checked
The sentence was found in the raw HTTP response body, read 2026-09-29, and the tool was then driven in a real browser with two test PDFs built for the purpose. Each carried a different marker string so that an upload could be detected by finding that text in a request body.
The Compare button only became enabled once both files were loaded, and the run produced the result quoted in the finding below. sourceHash is the SHA-256 of the raw response body, held as p13-logrunlab-compare-raw.html.
-
“You can compare two PDFs of up to 20 MB and 200 pages each.”
Checked by us - the source says this Read 29 September 2026
Both limits hold, and both are enforced at the point of loading rather than at the point of comparing. The 201 page file was refused with "This PDF exceeds the 200-page technical limit" and the 29.3 MB file with "This PDF exceeds the 20 MB technical limit", on both file slots, with the Compare button staying disabled.
The stated figures are therefore accurate and not rounded, since 201 and 29.3 MB are each just over the line. What is NOT established is where exactly the boundary sits below those figures, or whether the limits apply to the pair together rather than to each file.
Standing: Not re-read against the live source
Check it yourself: the source, the SHA-256, and how we checked
- From
- the raw HTTP response
- SHA-256
- 0947092a798846540e3fe59613b2c9a9c38df5484d583262a28ce5366e5b44b3
How we checked
Both limits were tested on 2026-09-29 rather than read. A purpose built 201 page PDF and a 29.3 MB PDF were loaded into the live tool.
Both limits are also stated twice on the page, once beside the file slots and once in the Limits and processing section. The source text was verified by raw HTTP fetch, held as p13-logrunlab-compare-raw.html.
-
“PDF Compare requires an existing usable text layer and does not run OCR. A mixed PDF can be accepted if some pages contain extractable text; image-only pages are not compared.”
Checked by us - the source says this Read 29 September 2026
Confirmed. The image-only file was refused on both slots with "No usable text layer was found.
Current version works only with text-based PDFs", and the Compare button stayed disabled. Refusing is the right behaviour and the message is unusually good: a tool that returned an empty comparison for a scanned page would leave the reader concluding the two documents were identical, which is worse than an error.
The boundary is what it says it is, and the limit is stated more precisely on the page than in the submission, which claimed only that it does not run OCR.
Standing: Not re-read against the live source
Check it yourself: the source, the SHA-256, and how we checked
- From
- the raw HTTP response
- SHA-256
- 0947092a798846540e3fe59613b2c9a9c38df5484d583262a28ce5366e5b44b3
How we checked
Tested on 2026-09-29 with a PDF built for the purpose: a single page whose only content is a JPEG and which contains no text operators at all, confirmed before use by reading it with a PDF parser (one page, total extracted text length zero) and by decoding the embedded image to check that legible text really is in the pixels. That last check matters, because an unreadable image would produce the same empty result from a tool that DOES run OCR.
The sentence quoted was verified by raw HTTP fetch, held as p13-logrunlab-compare-raw.html.
-
“PDF files, extracted text and comparison results stay in the current browser session and are not stored by Logrun Lab as document content.”
Checked by us - the source says this Read 29 September 2026
No document content left the device. Neither marker appears in any request, and the whole comparison ran in the tab.
Two POSTs were made to logrunlab.com/api/events carrying {"event":"pdf_compare_open"} and {"event":"pdf_compare_compare"}, 28 and 31 bytes, with no text from either document. Every request the page made, 25 in total, went to logrunlab.com, so no third party was contacted at all.
The claim about not storing content is one this test can support only in part: it establishes that nothing was sent anywhere, which is the strong reading, and cannot establish what a server would do with data it never received.
Standing: Not re-read against the live source
Check it yourself: the source, the SHA-256, and how we checked
- From
- the raw HTTP response
- SHA-256
- 0947092a798846540e3fe59613b2c9a9c38df5484d583262a28ce5366e5b44b3
How we checked
Two marked PDFs were put through the live tool on 2026-09-29 while every request was recorded. The marker check is the part that matters: a marker is a unique string in the file's text layer, so finding it in a request body proves the document itself was transmitted, and not finding it after the tool has demonstrably processed the file is evidence that it was not.
The instrument was proved before use with its own self test, whose three controls include a multipart POST carrying no file that must NOT be reported as an upload.
-
“Password-protected or encrypted PDFs are not supported in the current version.”
Checked by us - the source says this Read 29 September 2026
Confirmed. The file was refused on both slots with "Password-protected or encrypted PDFs are not supported in this version", and the Compare button stayed disabled.
It is detected and named rather than crashing or silently failing. The limit is stated on the vendor's own page as well as in their submission, so the disclosure and the behaviour agree.
Standing: Not re-read against the live source
Check it yourself: the source, the SHA-256, and how we checked
- From
- the raw HTTP response
- SHA-256
- 0947092a798846540e3fe59613b2c9a9c38df5484d583262a28ce5366e5b44b3
How we checked
Tested on 2026-09-29 with a PDF built for the purpose and encrypted at the simplest setting in the specification, /V 1 /R 2 with 40-bit RC4 and a user password, so that the file genuinely requires a password to open. The empty-password case was avoided deliberately: a document encrypted with an empty user password still opens freely, so testing that would have looked like a pass while testing nothing a reader would meet.
The sentence quoted was verified by raw HTTP fetch, held as p13-logrunlab-compare-raw.html.
Each claim is quoted from the vendor's own pages. This section describes what we have and have not checked - it is not a judgement about anyone.
Sponsorship is distribution, not an endorsement or a promise of search rankings, AI citations, traffic, or sales. This listing is a free entry, compiled from published sources.
Related reading
- Which PDF tools keep your file on your machine - what we count as processing on your own device.