How to make a Burning Bridges AI video

1. Choose the right Burning Bridges template

“Template” can mean three different things. Check what a page actually supplies before uploading a photo or paying.

This guide concerns reference-driven AI editing, not a downloadable CapCut project. Use a reference clip you are entitled to transform; a website example is not permission to reuse its footage or soundtrack.

2. Prepare one clear subject photo

For an adult portrait, use even light, an unobstructed face, and enough upper body to show the clothing you want. A tiny face or a crowded group image makes the intended lead harder to distinguish. Keep your original photo and reference video unchanged.

Note the reference’s length, orientation, frame rate, and audio timing. If you want the same performance with a different lead, avoid simultaneously requesting a new scene or different camera movement. Use photos and footage you have permission to transform. Our creation interface supplies a prepared motion reference, so you only choose your photo.

3. Describe the edit, not a substitute for the reference

This original reference instruction is for a consenting adult lead. Select and copy it into a compatible editor that accepts both a reference video and an identity image. It is not a validated prompt, an API request, or a guarantee of matching motion. Our public generator is not yet available to run it.

Use the supplied reference video as the guide for movement, timing, and camera motion.
Replace only the foreground lead with the consenting adult in the supplied portrait.
Keep supporting people in their existing roles; do not add a second lead.
Preserve the portrait's visible clothing and appearance throughout the shot.
Follow the reference hand gestures, including which hand holds the cup and where it moves.
Keep the cup visible and consistent when it is visible in the reference.
Follow the reference facial performance without inventing new dialogue.
Keep the scene fully clothed and non-sexual.
Create one continuous shot. Do not add cuts or a new camera move.

Adapt only the details relevant to your permitted material. The reference performance supplies the action. A prompt cannot force frame-by-frame gestures, preserve every prop, or lock lips to individual sounds; longer instructions are not automatically better.

4. Inspect a short sample before a longer version

Start with a five-second section containing the gesture, close-up, or prop interaction that matters most. A good first frame does not establish that the whole edit works.

• Replacement: is the intended lead replaced, rather than still visible behind the new subject?
• Hands and props: does the cup stay in the correct hand, and do the hand’s direction and reach follow the reference?
• Continuity: do the face, clothing, body, and surrounding scene remain coherent throughout?
• Mouth movement: compare the visual performance with the source, then listen to the export. The same soundtrack is not proof of the same lips.

Only consider a longer continuous version after the sample meets your needs. Our maximum is 15 seconds. A playable file can still fail these visual checks; technical completion is not a promise of perfect creative results.

5. Add music after the visual edit

Place the finished video and permitted audio on the same editing timeline. Align the same starting moment, trim the sound to the video, and preview both ends. If the recording is a little shorter, a brief silent tail preserves its speed; looping or stretching it changes the performance.

Adding music does not repair lip sync. It changes the soundtrack, not the rendered mouth shapes. If your song segment and reference performance have different timing, moving the audio alone may not make them agree.

The product target includes sound, but the public generation-and-audio workflow is still under validation. We do not supply a music license. The original-song explanation separates listening to the source from obtaining reuse rights.

When the result does not follow the reference

If the old lead remains, check whether the tool supports targeted replacement rather than general animation. If the cup disappears or hands change direction, more pixels alone will not fix the motion. If identities blend, simplify to one intended replacement. These are diagnostic starting points, not guaranteed remedies.

Keep the raw result before editing. Avoid repeating the same submission without understanding the failure. A provider error, an unreadable prompt check, an unsupported format, and an unattractive result are different outcomes; they are not automatically NSFW violations.

Availability, price, and prompt checks

One-time packs start at $4.99 USD for 500 credits, with no subscription. One generation uses 500 credits; browser audio assembly costs no extra credits. See pricing and credit returns versus payment refunds before purchasing.

AI checks the actual generation prompt for NSFW content, and provider safety filters remain enabled. This is not an AI review of uploaded images or finished videos, nor a verification of consent or copyright. Review any result before sharing.