getPNGgetSVG

Privacy

Your pictures are never uploaded. The background is removed by a model that runs inside your browser, on your own device, and no image data is sent to this site or to anyone else.

Your images

When you drop a photo, it is read by your browser and processed in a web worker on your machine. The site has no server that could receive it: the pages are static files. Close the tab and the picture is gone; nothing is stored on this device by the site either, apart from the model weights and a copy of the site's own pages, scripts and fonts in the browser's cache, which is what lets the tool open with no connection, and neither contains anything of yours. The one exception is a picture you send from one tool to the next with the Then menu: it waits in the browser's own storage for the moment the next page takes to open, and is deleted as that page takes it.

You can check this rather than trust it. Open your browser's developer tools, go to the Network tab, and remove a background. After your first picture, the only requests are the page itself, its scripts and fonts, and one small counter request described below. Switch the browser to offline and the tool still works.

What the site does collect

The AI model

The first time you use the tool, your browser fetches the AI model from getpng.app itself and keeps it in its cache; on a graphics card a second, smaller model for hair is fetched the same way the first time a picture has a soft edge for it to work on. Your picture is not part of either request, and no other website is contacted: the page's security policy does not allow a connection to any host except this site and Google Analytics.

The HEIC decoder

Browsers other than Safari cannot open HEIC, the iPhone's photo format, so the first time you drop a HEIC photo your browser fetches a decoder from getpng.app (about 2 MB) and keeps it in its cache. The photo is decoded inside your browser like every other picture; the request for the decoder contains nothing of yours, and pages that never see a HEIC file never fetch it.

The PDF reader

A PDF page is a set of drawing instructions rather than a picture, so turning one into an image needs a renderer. The first time you drop a PDF, your browser fetches pdf.js from getpng.app (about 1.7 MB, with the standard fonts and character maps fetched only if a document needs them) and keeps it in its cache. The document is opened and drawn inside your browser, and so is the password of a locked one: there is no address on this site that could receive either.

The PDF editor

Merging, splitting and compressing a PDF means reading a document somebody else wrote and writing it back out, which the renderer above cannot do. The first time you use one of those three tools, your browser fetches pdf-lib from getpng.app (about 500 KB) and keeps it in its cache. Your documents are opened, rearranged and saved inside your browser.

The upscaling model

The AI engine on Upscale an image needs a trained model, so the first time you enlarge a picture with it your browser fetches the weights for that scale from getpng.app (about 900 KB) and the TensorFlow.js runtime that executes them, and keeps both in its cache. The model then runs on your own graphics chip: the picture is never sent anywhere, and the requests for the model contain nothing of yours. The other engine on that page needs no download at all.

Open source components

Under the GDPR

Because your images are never transmitted or stored, the site does not process them in the sense of the regulation: there is nothing to keep, nothing to delete and nothing to hand over. The daily counts are totals per day (removals, and how many used each tool, format or document) and cannot be related to a person. Google Analytics is used under Google's standard terms, and you can block it with any content blocker without affecting the tool.

Contact

Questions about any of this: hello@getpng.app.