Burning Bridges AI Photo Guide & Upload Help

Check the file before you upload

The studio accepts one photo of one consenting adult per video. Use an original JPG, PNG or WebP image that meets all of these requirements:

• File size: up to 10 MB.
• Short edge: at least 340 pixels.
• Shape: the longest edge must be no more than 2.5 times the shortest.

For example, a 900 × 1200 portrait meets the dimension checks. A 200 × 300 thumbnail is too small; a 400 × 1600 strip is too narrow. Renaming a HEIC file to .jpg does not convert it. Export unsupported formats to a real JPG, PNG or WebP first.

The browser resizes the longest edge to at most 1600 pixels and converts the upload to JPEG before submission. A very narrow image can become too small on its short edge after that resize. Start with a balanced portrait, not a panoramic crop.

A face the model can read

Choose a sharp, well-lit face with the head fully inside the frame. A front-facing or gentle three-quarter view is a practical starting point. An upper-body portrait can also show the neckline and outfit that the reference performance uses.

Avoid sunglasses, masks, heavy beauty filters, hair covering the eyes and a face that occupies only a tiny part of the picture. Keep only the intended adult in the photo. A group image does not tell this single-person workflow which friend you mean.

You can crop away empty space in a photo editor before uploading. Preserve enough of the head and upper body to describe the person clearly, and avoid stretching the face to force a different shape.

What to do when an upload is rejected

Unsupported file: export it as JPG, PNG or WebP and try the exported file. The filename extension alone is not enough.

File too large: reduce its dimensions or export a compressed JPG under 10 MB. Use a version where the face still looks sharp at normal viewing size.

Image too small or too narrow: choose the original camera photo instead of a messaging thumbnail, or make a less extreme crop. Enlarging a tiny screenshot adds pixels but cannot recover missing face detail.

File cannot be opened: confirm it opens in your usual photo viewer, then export a fresh copy. If a content check rejects a photo, use a different permitted image; changing the file extension does not address that check.

If the face or hands change in the result

Watch the beginning, the turns and the hand-over-eyes gesture. Occlusion and motion can make a face less consistent even with a valid upload. Compare your input and result: is the face too small, badly lit or hidden in the photo? For a future attempt, change the input issue you identified rather than assuming a higher-resolution file alone solves it.

A completed video with a creative detail you dislike is different from a confirmed failed task. The studio returns credits for confirmed failed or canceled jobs; it does not promise a free retry for every likeness difference. See pricing and credit rules.

Photo privacy and drafts

Your photo is processed by cloud AI services. The current workflow does not save the input portrait to the task database or R2, but the provider handles its own retention. Avoid confidential images and read the privacy policy.

The browser may keep a draft so you can return after signing in or buying credits. It expires 24 hours after the last save and is discarded on your next studio visit. Use Clear photo draft to remove it immediately, especially on a shared device.

Open the studio when your photo is ready, or follow the full video tutorial.