Last term I had to submit internship paperwork. The system took JPG only, 2MB per file maximum. My phone shoots HEIC at four or five megabytes each, so I failed both conditions at once. That evening in my dorm I searched for how to convert HEIC and everything I found either wanted an account or capped me at a few files a day, which a student budget doesn't stretch to. I ended up on ImgIng, converted and compressed in one pass, no signup. It's at imging.ai.
A semester on, I think a student needs exactly three things from an image tool: it costs nothing, it doesn't want an account, and the file doesn't get sent somewhere you can't see. That sounds easy. Very few manage all three.
What I used it for
Paperwork was the start. After that: a batch of coursework images with faces in them that I'd rather not send anywhere; pulling figures out of paper PDFs for a lab meeting; cutting a transparent background for a poster; turning a few screenshots into a GIF for a report. It reads a lot of formats, so whatever odd file a classmate sends usually opens — and compression, format conversion, AI background removal, the animation workshop, OCR, an image editor and AI upscaling all live in one place, with PDF compression, PDF to image, pulling embedded originals straight out, content extraction and HTML both ways on the document side, plus DOCX and PPTX to offline HTML. I don't need most of that, but when I do I'm not off hunting for another site.
For the coursework batch I actually checked whether anything was being sent. Open DevTools, Network tab, process one image, look for a POST — there wasn't one, and the preview's src was a blob: URL, which is an object in the browser's own memory rather than anything that crossed the network. Took under a minute and told me more than a privacy policy would. The lab-meeting one is worth mentioning too: I started by converting whole PDF pages to images and cropping, and only after a while noticed there's an "extract images" function that pulls the embedded originals out directly, at better quality. Converting a page and extracting an image don't give you the same thing — I had no idea at the time.
The annoying parts
Background removal downloads a model the first time, and my ageing laptop took a while over it; the fan spun up too, because it's my own machine doing the computation. A few formats also go to their server rather than running locally, colour-coded in the UI, so "not uploaded" only covers the features that stay local.
One thing I still haven't worked out: compression picks its own parameters, and a screenshot gets a different profile than a photo. I tried to understand what it decides on, read the explanation on the page, and didn't fully follow it — all I know is the results have been consistently fine and I've never adjusted anything by hand. For now I use it as a black box and I'll figure it out eventually. A semester in, I haven't paid for image processing or created an account, and for a student those two are most of what matters.
