Bottom Line Up Front (BLUF): Colorization adds colour to old black-and-white photos right in your browser. Upload a picture, press the button and get a colour version: you can compare it with the original on the page and download it as an image. The photo is never uploaded, everything runs on your device, no tokens are spent and there is no paid API here.
It adds colour to a black-and-white picture. The input is a photo, the output is the same photo in colour. It helps when you are sorting a family archive, preparing an old portrait for printing, framing a picture from a trip, or showing a city's history on old shots.
A colour version brings a frame to life that otherwise reads like a document: clothes, sky, greenery and faces are easier to make out in colour.
The photo is scaled down to 256 by 256 pixels and the model predicts a colour for every pixel at once — it sees the whole frame, not separate fragments. The colour image is then stretched back to the size of the original.
The result is delivered as a PNG image: you can compare it with the original on the page and download it.
The photo never leaves the browser: the model weights are downloaded once from an open repository and cached by the browser, then the computation runs on your device. The tool has no paid API, so no tokens are spent.
The first run takes longer — that is the weights download. Later colorizations are faster because the model is already cached.
Close-ups work best: faces, clothes, sky, grass, water, house walls. The picture should be sharp — on a very noisy frame the colour lands in patches.
This is artistic colorization, not a restoration of the real colours: the model guesses the shade of a dress or a sign. Small details, text and jewellery may get an inaccurate colour.
Pick a picture with a large subject and little grain: the more recognizable shapes in the frame, the more confident the model is about colour.
Compare the result with the original on the page: if a shade looks wrong, try another frame from the same series — colorization depends on what is in the frame. The finished file downloads as PNG.
Yes. The model weights are loaded into the browser and the computation runs on your device. Neither the photo nor the result is sent anywhere — the tool has no paid API at all.
Nothing. No tokens are spent on this tool: it does not call paid services, everything runs on your device.
Images: PNG, JPG, JPEG, WEBP and other common formats. The tool does not accept video.
On the first run the browser downloads the model weights — about 61 MB. After that they come from the cache and colorization is faster.
No, this is artistic colorization: the model picks plausible shades but does not know the true colour of an object. Small details and text may get an inaccurate colour.
No. The tool works with one photo at a time. The section has separate tools for video.