Best Photo for a Burning Bridges AI Video

A Burning Bridges AI video is only as good as the photo that goes in. The model has one picture of you and has to rebuild your face in every shot — at the head of the table, leaning into the camera, hand over your eyes, turning left and right. Anything it can't see in the photo, it has to guess, and guesses are where "that doesn't look like me" comes from.

The good news: the photo that works best is an ordinary one. You don't need a photoshoot.

The quick checklist

• One face in the photo. Just you, nobody else in the frame.
• Facing the camera, both eyes visible. Straight on or a slight turn, not a side profile.
• Even, soft light. Daylight from a window is ideal. No harsh shadow across half your face.
• Head and shoulders. Fill most of the frame with your face and upper body, not a full-body shot from across the room.
• Sharp and at least 720px wide. The original file from your phone, not a screenshot.
• Your normal look. A recent photo, without a beauty filter.

If you have a photo you'd use as a profile picture, that's usually the one.

Why each of these matters

Eyes. The trend's signature move is a flat hand held over the eyes, then a slow look around the room. If your eyes are hidden by sunglasses, a hat brim or glare on your glasses, the model invents them — and the face drifts from shot to shot.

Light. The party scene has its own warm chandelier light, so the model relights your face anyway. It can do that well from an evenly lit photo; it struggles with a face that's half in deep shadow, because it can't tell where your features end and the shadow begins.

Distance. In a full-body photo your face might be 80 pixels wide. That's not enough detail to rebuild it across several close-ups. Crop in, or pick a closer photo.

Filters. Beauty filters smooth skin, enlarge eyes and reshape jaws. The model copies what it sees, so a filtered photo gives you a filtered-looking video — often uncanny once it starts moving.

Clothes. You don't need to dress up. In our generator your outfit is restyled into a sharp going-out look in the spirit of what you're wearing, so a T-shirt photo is fine.

Photos to avoid

• Group photos, or a friend cropped half out — extra faces and shoulders make it unclear who the host is.
• Sunglasses, masks, hats over the eyes — the model guesses your eyes and the likeness slips.
• Heavy filters or face-editing apps — the video inherits the filtered, plastic look.
• Very close, wide-angle selfies — the lens stretches your nose and face, and the AI keeps the distortion.
• Tiny, blurry or screenshot-of-a-screenshot images — not enough detail to rebuild your face in close-ups.
• Children, celebrities, or anyone who hasn't agreed — not allowed. Use photos of yourself or people who said yes.

A 10-second test before you upload

Open the photo and zoom in on your face. If you can clearly see both eyes, the edges of your hair and your mouth, it's a good input. If it gets soft or blocky as soon as you zoom, pick another one.

Making it with a friend

Use two separate photos, one face each, not one photo of you together. Try to pick photos with similar light and sharpness, so one of you doesn't look noticeably better than the other. The full walkthrough is in How to make the Burning Bridges AI video with a friend.

Making it with your pet

The same rules apply, at your pet's eye level:

• One animal, face fully in frame.
• Get down to its height and shoot straight on, not from above.
• Markings and fur colour clearly visible in good light — the model keeps the breed, fur and markings from the photo.
• Sharp, not mid-jump. A sleepy, still pet beats an action shot.

If the video doesn't look like you

AI likeness varies from render to render, even with the same photo. If a result misses, the fix that helps most is a better photo: closer, front-facing, in daylight. If a render fails outright on our generator, the video goes back to your account automatically so you can try again.

Making it yourself with a prompt

Everything above applies just as much if you use Seedance, Kling or another image-to-video model directly. Pair a good photo with our free Burning Bridges AI prompt, or read what the Burning Bridges AI trend is first.

File types: our generator takes JPG, PNG or WebP up to 15 MB.