Most AI image tools start from a blank box and a text prompt. That works for scenes invented from scratch. It falls apart when you already have a photo — a selfie, a landscape, an old family print — and only want to change the style, the hair, or the damage on the print.

That second job is image-to-image . The model starts from your picture, keeps the person, the pose, and the composition, and changes only what you ask for. Below is a plain guide to how that works, plus a step-by-step run on image2img.net , a free browser tool built for exactly this.

What image-to-image actually does

Text-to-image tools invent a scene from words. Image-to-image tools take an existing photo as the base. You keep the identity and layout; you steer the look.

Common uses:

Restyle a selfie (anime, film, collage, doodle)

Preview hairstyles on the same face

Repair scratches and fading on old prints, then add natural color

Turn a casual selfie into a clean studio-style headshot

You do not need to write long prompts.…