How to choose photos for an AI video
Clear subjects, the right number of photos and permission to use them: a short checklist before you upload.
Start with the template’s photo slots
Each template explains how many photos it needs. Burning Bridges currently prepares a single-photo draft; the hug draft has two slots. Use one reference per requested subject and keep each person in the intended slot. A group picture does not automatically replace two separate inputs.
Make the subject easy to recognize
Choose an evenly lit, sharp photo with a visible face for portrait effects. Avoid tiny faces, sunglasses, heavy filters or objects covering important features. Keep enough body visible when the intended action needs it. Good input improves the starting reference; it cannot guarantee identity or choreography.
Check the file before uploading
The current photo form accepts JPG and PNG, up to 10 MB, at least 256 pixels on each side and no more than 4 megapixels. Resize a larger image before choosing it. Renaming a file does not change its format. Browser checks are preliminary; live uploads also need independent server validation.
Use photos you have permission to use
Use your own photos or get permission from the people shown. Do not create deceptive impersonation, harassment or intimate content. Follow the template’s age and content rules. In the current guest preview, photos stay in the tab; saved drafts contain text and settings, so photos must be selected again after reopening.