1. Your photos are prepared in the browser
Crop and rotation run on your device. Only the adjusted versions are sent; each photo can be up to 10 MB.
2. Your choices become instructions
Blendly turns your style, strength and prompt into a written instruction for the AI. Strength picks one of three levels: subtle (below 34), noticeable (34–66) or bold (67 and up). Pose styles such as Hug or Handshake add pose directions; multi-photo blends ask for one shared scene, never a collage. Your own prompt is added at the end, up to 1,000 characters.
3. An AI image model creates a new picture
The photos and instruction go to an image-editing AI model through Lovable AI Gateway. The model does not cut and paste pixels; it paints a new image using your photos as references. That is why faces and small details may change slightly.
4. The image streams back
You see a blurry preview while the image forms, then the finished picture. The canvas is 1024×1024 for square, 1536×1024 for landscape or 1024×1536 for portrait.
5. Saving and downloading
If you are signed in, the finished image and its settings are saved to your private gallery, visible only to you, and kept until you delete them. Downloads are converted to PNG, JPG or WebP in your browser.
