Get the Best Results: Products, Prompts and Details
What to upload, how much to write, and how to keep small product details right in your Fittins AI photos and videos.
Fittins AI already does most of the writing for you. When you press Generate, it looks at your product photos and the model you chose, describes them in detail, and builds the full instructions for the AI. Your part is small: the right photos, a short prompt, and a few words about any detail that must not be missed.
Product photos: angles yes, colors no
You can add up to 3 product photos. The model wears what those photos show.
- Different photos of one item (front, back, a close-up of the logo): good. They help the AI get the item right.
- Items worn together (a top, trousers and shoes): good. The model wears all of them.
- The same item in different colors: make a separate generation for each color. If you add a grey, a teal and a white hoodie together, the AI tries to put all three on the model at once.
Example
Selling a hoodie in 3 colors? Generate the grey one, then the teal one, then the white one, with the same model and prompt each time, so the photos match.
Keep the prompt short
The person comes from the model you pick, and the product's design comes from your photos. So the prompt only needs what they can't show: the scene, the pose and the mood. A few words are enough.
Good prompts
- walking in a park, smiling
- standing in a bright studio, hands in pockets
- sitting at a café table, looking at the camera
Don't paste long prompts from ChatGPT
A long prompt that also describes the person, the clothes, the camera or the video length competes with what Fittins AI already knows from your photos and settings, and it can confuse the AI. If you started in ChatGPT, keep only the scene and the pose.
Mention small details that must stay
Some details are easy to miss in a photo: a second button, a small logo, a stitch line. Name them in the prompt in a few words, and the AI pays attention to them.
Examples
- two buttons on the cap
- small logo on the left boot
- white stitching on the pocket
Scene first, then the detail
"walking in a park, smiling, two buttons on the cap"
Use the controls for settings
The model, the aspect ratio and, for videos, the length (5 or 10 seconds) are chosen with the controls on the generate page. Writing them in the prompt won't change them.
Quick checklist
Photos: angles of one item or items worn together; one color per generation. Prompt: a few words about the scene and pose. Details: name the small thing that must stay. Settings: use the controls.
Continue reading
