Compare documents without uploading them
If you have landed here, the question has probably already been settled for you: the document is a contract, a draft under negotiation, an employment file, a medical report or a tender, and uploading it to a website is not something you are willing or permitted to do.
This is a comparison that has no upload. Not an upload with a deletion policy, not an upload to a region of your choosing, not an upload over an encrypted connection: no upload endpoint on the site at all. Both documents are read by the browser on your own machine, compared there, and displayed there. The sections below explain what that means technically and how to verify it yourself in about thirty seconds.
or drop it here
.pdf · .docx · .txt · .md
or drop it here
.pdf · .docx · .txt · .md
Nothing is uploaded. Both files are read inside this tab.
How to do it
- 01
Open the network inspector first, if you like
Press F12, open the Network tab, then compare two documents. You will see the page and its code load, and no request carrying document data. This is the only verification that means anything, and it takes less time than reading a privacy policy.
- 02
Drop both files
They are read through the browser’s own file API, which hands the bytes to the page without sending them anywhere. The same mechanism that lets a photo editor open a photo.
- 03
Work offline if you want to
Once the page has loaded, the comparison needs no network at all. Turn the Wi-Fi off and it still works, which is itself a demonstration that nothing is being sent.
- 04
Nothing is stored either
Closing the tab is the end of it. The documents are never written to disk by this site, never placed in local storage and never cached; reopening the page gives you an empty pair of slots.
What uploading a draft contract actually commits you to
The risk is usually described as "the company might read your file", which is the least of it and the easiest to promise against. The real exposure is in the places a document goes once it has been accepted by a server, none of which appear in the interface.
It is written to disk, at least temporarily, on a machine you do not control. It is very likely written to more than one, because the service is load-balanced or replicated. It appears in request logs, and logs are retained on a different schedule from files and are frequently forwarded to a third-party log aggregator. It may pass through a CDN. It becomes part of that company’s backup set, which means a deletion policy measured in hours is bounded below by a backup retention measured in weeks. And the whole of it is now within the scope of any breach that company suffers, in the future, for reasons nothing to do with you.
For a holiday photograph none of that matters. For a draft agreement it is a disclosure to a third party, and for a document covered by professional privilege, by a confidentiality undertaking in another contract, or by a duty of medical or legal confidence, it may be a disclosure you were not entitled to make. The question stops being about trust and becomes about authority.
A browser-based comparison removes the question rather than answering it. The bytes are handed from the file system to the page by the browser itself and never enter a request; there is no endpoint on this domain that accepts a document, so there is nothing to retain, nothing to log, nothing to back up and nothing to be in the scope of a breach. The strength of that is not that the promise is better worded: it is that there is no promise, only an architecture in which the thing cannot happen.
Questions
- How can a web page compare documents with no server?
- The same way a browser shows you a PDF: the code runs on your machine. The PDF parser, the Word unzipper and the comparison engine are all JavaScript delivered to the tab, and they read the bytes the browser hands them. The site serves the program; your computer does the work.
- How do I verify it rather than take your word for it?
- Open the browser’s network inspector before comparing, and watch. You will see no request carrying document data. Or simpler: load the page, disconnect from the network, and compare two documents. If it still works, nothing is being sent.
- What data does the site collect at all?
- Anonymous page analytics and a small, fixed list of interaction events: that a comparison ran, roughly how long it took, which format pair was used, which view was chosen. Every value is a word or a range chosen in advance, with no free text, so no file name or document content can reach them even by accident. The privacy page lists every event.
- Are the documents stored in my browser?
- No. Nothing is written to local storage, IndexedDB or a cache. Closing the tab ends it, and reopening the page gives you empty slots. The one thing deliberately kept nowhere is your documents.
- Does it work on a locked-down work computer?
- Usually yes, because nothing is installed and no administrator rights are needed: it is a web page. Whether your organisation permits document comparison on a web page at all is a question for them, and the fact that nothing is transmitted is the part worth putting in front of them.