Why You Should Enhance Photos Without Uploading Them
Published
Most online photo enhancers work the same way: you upload your image, a remote server processes it, and you download the result. That is convenient, but it means a copy of your photo now lives on someone else’s computer.

What Happens When You Upload a Photo
When an image is sent to a cloud enhancer, you usually cannot verify:
- How long it is stored and whether it is really deleted.
- Who can access it, including staff, contractors or third-party AI providers.
- Whether it is used to train models.
- Where the servers are, and which privacy laws apply.
For a holiday snapshot that may be fine. For ID cards, contracts, medical images, children’s photos or internal screenshots, it is a real risk.
How In-Browser AI Works
Modern browsers can run neural networks directly on your hardware:
- WebGPU gives web pages access to your graphics card for fast AI inference.
- WebAssembly (WASM) runs compiled code at near-native speed when WebGPU is not available.
- ONNX Runtime Web loads the AI model, here Real-ESRGAN, into the page.
The model file is downloaded once, like any other web asset. After that, your image is decoded, enhanced and encoded entirely in the page’s memory. Nothing about the image is sent back.
How to Check It Yourself
You do not have to take our word for it:
- Open Crispix and wait for the page and AI model to load.
- Open your browser’s developer tools and switch to the Network tab.
- Enhance an image.
- Notice that no request carries your image to a server.
You can even disconnect from the internet after the model has loaded and the tool keeps working.
Trade-offs to Know
| Cloud enhancer | In-browser (Crispix) | |
|---|---|---|
| Privacy | Image leaves your device | Image stays on your device |
| Speed | Depends on upload and queue | Depends on your GPU / CPU |
| Account, limits, watermark | Often required | None |
| Very large images | Server can handle | Limited by device memory |
On older phones or very large photos, in-browser processing can be slower. For most everyday images, a modern laptop or phone finishes in seconds.