Pixel2Page

Drop in a screenshot of a website or app. SmolVLM, a tiny vision language model, reads it on your device while a layout analyser measures sections and colours. Together they produce editable HTML and CSS.

Click or drop a screenshot
PNG or JPG, stays on your device
Model: SmolVLM 256M Instruct (about 250 MB, cached after first run).

Detected palette

What the model sees

Waiting for an image.