Which Tools Work Offline

Every tool, and what it actually does

This table is not written by hand. It is generated from the two lists in the URL contract that the code itself reads, so a tool cannot be described here as doing one thing while doing another. A tool that gains a dynamic import changes this page in the same commit that changes the code.

Nothing is uploaded and works offline are different claims

The first says your file is not transmitted. It is true of twelve of the fifteen tools here, and you can check it in about twenty seconds by opening the network panel and watching that no request carries your file.

The second says the tool would still function with no connection at all. That is stronger, and it holds only where the page already has everything it needs before you drop a file. Most of these tools deliberately do not load their engine until a file arrives, because loading a PDF library or an image codec on every page view would make the site slow for everybody who never opens that tool.

Both claims can be honest. Only one of them can be made about every local tool, and this site got that wrong once: nine pages said some version of disconnect and it still works, before anybody had measured it.

How to check the upload claim yourself

Open the developer tools in your browser and go to the network tab. Clear it. Drop a file into any tool on this site and watch what appears.

You will see requests for code the page needs, which are the dynamic imports described above. What you will not see, on any of the twelve, is a request carrying your file. On the three that do upload, you will see exactly that, because that is what they do.

This is the check worth learning, because it works on any site rather than only this one, and it does not require believing anybody.

How to check the offline claim yourself

Load the page first. Then turn off your wifi, or set the network tab to offline. Then drop the file in.

The order matters. Cutting the connection before loading the page proves nothing except that pages have to load. The claim is about what happens once the page is already open.

Why three tools upload, said plainly

The background remover and the two transcribers do work a browser cannot do well enough to be worth shipping. Those files go to our servers, get processed, and are dropped: not stored, not logged, and never used to train anything. For video, the audio track is extracted on your machine and only that is sent, so the video itself never leaves.

The reason to say this in plain words is that the alternative is what the previous version of this site did, which was to tell people their images never left the device while posting them to an API. A privacy claim that is not true is worse than no claim at all, because it is the one thing a reader cannot check without effort, and therefore the one thing most worth being honest about.

What this page is for

It is not here to rank for anything. Almost nobody searches for this, which was measured rather than assumed. It exists because every other page on this site makes a claim about where your file goes, and one page ought to gather those claims in a single place, next to the instructions for disproving them.

More guides

• What Is a WebP FileThe format you did not choose, saved from a web page, and now something will not open it.
• AVIF vs WebPAVIF wins on size. Size is not the only thing you are choosing between.
• JPEG vs PNGOne rule decides almost every case. The exceptions are the part worth reading.

FAQ

Do these tools upload my files?

Twelve of the fifteen do not. Three do: the background remover and the two transcribers, and the video one sends only the audio track. The table above says which is which, and it is generated from the code rather than typed out.

How can I verify that nothing is uploaded?

Open your browser network panel, clear it, then use the tool. You will see the page fetch code it needs, and on the twelve local tools you will not see any request carrying your file.

Which tools work offline?

The three whose work is plain canvas already sitting in the page: resizing, adding a background colour, and blurring or censoring. The others fetch an engine when a file arrives, so they need a connection at that moment even though they never send your file anywhere.

Why do the other local tools need a connection?

They load their engine, a PDF library or an image codec, at the moment a file arrives rather than on every page view. That keeps the site fast for visitors who never open that tool, and the cost is that a first use needs a connection.

What happens to the files that are uploaded?

They are processed and dropped. Not stored, not logged, and never used to train anything. There is no database in this application to keep a copy in.

Is any of this checked automatically?

Yes. A test in the build refuses to ship a page claiming to work offline unless the tool it describes is in the list the code reads, and the table on this page is generated from that same list.