Image-to-Image

Give it a reference image plus a description, and AI generates a new version. Common uses: style transfer (realistic → watercolor / oil painting / anime), detail edits (new background, hair color, accessories), pose variants of the same person, product photo variants in different colors, and AI avatars (turning a regular photo into illustration style).

Yes. Your culling picks, AI enhancement results, and previously generated images all land in your local library automatically — just click "Library" to pick one. You can also drag-and-drop or upload a new image (up to 20MB). A clearer reference with a well-defined composition gives better results.

Yes. Upload a clear front-facing or half-body photo, and write a prompt like "cartoon illustration style / oil painting portrait / cyberpunk, keep the facial features," pick a 1:1 square ratio, and the result works great as a profile picture.

Very small, blurry, poorly lit, or cluttered images give the model less detail to work with. Start with a clearer reference when possible. Requests are also subject to the selected model's content restrictions.

You can upload a JPEG, PNG, or WebP file up to 20MB, or pick any image already in your local library. For the cleanest result, use a sharp reference of at least 512x512. Very small or heavily compressed images give the AI less to read and produce weaker output.

Yes. Upload the photo and describe the look, for example "anime style, keep the facial features" or "flat cartoon illustration." If you would rather pick a preset than write a prompt, the AI Stylization tool covers Ghibli, anime, oil painting, and watercolor in one click.