How Stretching Stays on Your Device

Stretching an image is arithmetic on pixels you already have. Width, height, and the color of each pixel are enough. That work fits in the browser you already opened. ImageStretcher draws the preview on a canvas in the page, and the download writes a file to your device. The original you dropped in is not stored as part of that stretch.

This page is the map of what stays put and what leaves. The legal wording, including cookies and advertising, is the privacy policy. Nothing here adds a collection the policy does not already name.

What the page actually uses

You open a file with the picker, a drop, or a paste. The browser hands the page an in-memory image. The stretcher reads its width and height, computes the output size from the numbers you type or the template you pick, and paints the remapped picture into a canvas. When you download, the canvas is encoded to PNG, JPEG, or WebP and saved through the browser’s download. Refresh the page, or leave it, and the preview is gone unless you already saved that download.

No account is required for stretch, crop, compress, or convert. The same canvas path is what crop image, compress image, and convert image use. Convert keeps the pixel size and changes the container. Compress hunts a quality for a target file size. Neither one needs a server to multiply or requantize pixels.

Actions that do send something

Four paths leave the device, and each one is a choice you can see:

Action	What is sent	What we keep
Stretch, crop, compress, convert, download	Nothing for the image itself	The file stays on your device. We do not store the picture.
Paste or type a public image URL	The page asks our server to fetch that address once, because the browser cannot always read another site’s image	The file and the URL are not stored. After it loads, stretch still runs locally.
AI Image Expander, Remove Object, Remove Text	A temporary smaller copy, and for the erase tools the area you brushed	Used for that request. Not kept on our servers afterward. Usage counters may be stored to limit abuse.
Share link	A copy of the result with a small ImageStretcher.io mark	A public page at that link. Shared images stay available; there is no scheduled deletion. Skip the button and nothing is uploaded.

Sign-in, feedback, and payment are account features. They store the account data the privacy policy lists. They are not required to stretch a file. Buying Pro or credits goes through Stripe. We store a customer id and subscription status so credits can be applied. We do not store the full card number.

Why a canvas is enough for a stretch

A stretch does not need to “understand” the photo. It maps each output pixel back to a place in the input and samples a color. That is a local loop. A server would only add a copy of a file the browser already decoded. Keeping the loop in the page means a client proof, a half-finished layout, or a personal photo can change size without a second copy sitting on someone else’s disk.

The limit is the device. A very large photo, or a Batch queue of large print sizes, uses the memory the browser will give the tab. If the preview stalls, use a smaller source or a smaller output. There is no upload quota on stretch, because there is no upload. Batch still caps a single queue at 20 files so one tab does not try to hold an unbounded folder. That cap is described in how to stretch multiple images at once.

What “we never upload your file” is claiming

On a stretch landing, that sentence means the stretch itself. It does not mean every button on the site is local. Read it next to the table above.

• Download is local.
• Share link is not local. The button says it uploads a marked copy.
• An image URL is a one-time fetch of an address you typed. Your local file was not that address.
• AI expand and the erase tools are not local. They exist because filling new pixels, or filling a hole, is a different operation from remapping the pixels you have. The pages say a temporary copy is sent when you generate.

If a workflow cannot accept any of those four exceptions, use the stretcher, do not paste a URL, do not generate, and do not press Share link. The download is the whole product for that job.

Cookies are a separate channel

Analytics and, when ads run, advertising cookies are about the visit, not about the picture in the canvas. Google Analytics measures visits. Microsoft Clarity loads only after you choose Allow analytics. Advertising, including interest-based ads and the consent message in the EEA, the UK, and Switzerland, is described in the privacy policy. Those cookies are not how the stretcher reads your file. An ad request is not a place we put your name, email, or phone number, and the page does not ask the browser for a precise location.

How to confirm it while you work

You can watch the stretch without taking the policy on faith.

1. Open the home tool with the browser’s network panel visible.
2. Drop a local file. Stretch, change the size, download. The image bytes of that file should not appear as an upload to ImageStretcher. The page already had them.
3. Press nothing else. Leave the page. The preview is gone. Your download, if you saved one, is the copy you have.
4. If you then use Share link or AI expand, you should see a request. That request is the exception you opted into.

Paste of a screenshot is the same local open, with the clipboard as the source. The steps are in paste a screenshot and stretch it.

What this does not prove

A local canvas does not make the picture safe from the device you are using. Someone with access to that computer can still open the download. A workplace browser policy can still inspect traffic for the requests you do choose to send. The claim is narrower: stretching, cropping, compressing, converting, and downloading do not require us to receive the file.

For the full list of stored items — accounts, votes, feedback, shared images — use the privacy policy. For the stretch itself, open the tool and keep the file on the machine that already has it.