How to make a Burning Bridges AI video

Every route that looks like the original works the same way underneath: the movement comes from the real party scene in Drake’s music video, and the face comes from a photo. So the do-it-yourself routes need three things.

1. A motion reference: a short clip of the party scene, trimmed to the part you want (the walk, the hand over the eyes and the scan of the room). About ten seconds is the usual length.
2. A clear photo of the person: one person, front-facing, whole face visible, head-and-shoulders or half body. The photo guide explains why each of these matters.
3. A motion-transfer tool, with credits or a subscription on your account.

You also need consent. Use your own photo, or one from someone who agreed to be in the video, and never a photo of a child. The music video belongs to its owners, so keep your version a short personal meme and do not pass it off as official.

Higgsfield Genjutsu is one of the motion-transfer tools the viral versions were made with. It copies the motion of a driving video onto a person from a reference image. Its menus change often, so these steps describe the workflow rather than exact buttons.

1. 01

   Trim the clip

   Cut the party scene to about 10 seconds, starting a second before the hand goes up. Export it as a standard MP4.

2. 02

   Upload your photo

   Add a clear, front-facing photo as the reference. If the tool takes more than one, add a second photo of the same person from a slightly different angle.

3. 03

   Add the driving clip

   Upload the trimmed scene as the motion source and pick the aspect ratio you plan to post in.

4. 04

   Generate

   Run it. If there is an instruction field, ask it to keep the person’s identity and the original room and lighting.

5. 05

   Review

   Watch the face during the hand-over-the-eyes moment and in the last second, where drift shows first. Try another take if it slips.

6. 06

   Export

   Download the best take. If the output has no audio, add the song when you post.

The prompt guide has copyable instruction text for that field. Higgsfield bills through credits and plans that change, so check its pricing page for what a clip of this length costs before you generate several takes.

Kling’s motion control does the same job: a reference image of the person plus a video whose motion it should follow. The workflow is close to Genjutsu’s.

1. Open Kling’s motion control mode (the name can change between versions).
2. Upload the reference image of the person.
3. Upload the trimmed party scene as the motion video.
4. If there is a text field, add a short instruction: keep the face from the image, keep the room and the lighting from the video.
5. Generate two or three takes and keep the one where the face holds steady through the gesture.

Kling bills per second of output or through a subscription, depending on the plan. A ten-second clip costs more than a five-second one, so test on a short trim before you run the full length. Check Kling’s pricing page for current numbers.

CapCut has user-made templates for trends like this one. It is the fastest do-it-yourself route on a phone and needs no driving clip, because the template already has the scene set up.

1. In CapCut, search the templates for “Burning Bridges” or “Drake AI”.
2. Watch the previews of a few. Pick one where the swapped face looks clean through the gesture, not just the first result.
3. Use the template and pick your photo when it asks.
4. Let it process, check the result, then export it or post it directly.

Results vary a lot. Some templates use real motion transfer; others lay a face over the lead that slides when the head turns or the hand crosses it. Templates also get removed, so one you saw yesterday may be gone today.

• The face drifts: it looks like you at the start and like a stranger by the end. Use a sharper, front-facing photo, add a second reference photo if the tool allows, and keep the clip short. Longer clips drift more.
• The hand over the eyes confuses the model. When the hand covers part of the face, some tools lose track and redraw a different face after it. Start the clip a second earlier so the model locks onto the face first, or try another take.
• Lighting mismatch. A bright daylight selfie in a dim, chandelier-lit room looks pasted in. Pick a photo in soft, even light without flash or strong colored light.
• Small or cropped faces. A face that fills only a corner of the photo, or one cut off at the forehead, gives the model too little to work with. Use a head-and-shoulders or half-body photo.
• Smudges around the eyes. Sunglasses, hands or hair over the face in the photo often turn into blur. Use a photo with the whole face visible.
• Hair, skin tone and expression change. Models tend to borrow from the original scene, so hair can come out shorter, skin tone lighter and a smile can disappear. A neutral, evenly lit photo keeps closer to the real person.

• Crop for the platform. TikTok and Reels are vertical (9:16) and the music video is wide, so either generate in 9:16 or crop around the lead and keep the hand gesture inside the frame.
• Use the official sound. On TikTok and Reels, pick “Burning Bridges” from the sound library instead of relying only on audio in the file. That links your post to the trend’s sound page.
• Keep the caption short. The clip does the work; “where she at?” or the name of who is in it is enough.
• Label it as AI where the platform asks, and tag people only if they agreed to be in it.

If you just want the clip, Burning Bridges AI Video skips the driving clip, the tool and the retakes. It uses the original party scene as the motion reference, so you only bring the photo.

1. 01

   Upload one photo

   Pick a clear, front-facing photo of one person with the whole face visible. JPG, PNG or WebP up to 10 MB.

2. 02

   Sign in and make it

   Sign in with Google and use 1 credit; making the video confirms the photo is you or that you have the person’s consent. Credit packs start at $4.99, with no subscription.

3. 03

   Download your video

   The AI puts you in the Burning Bridges party scene in a few minutes. Download the MP4 and post it to TikTok, Reels or X.

The result is a 720p video of about 10 seconds with the song audio. Credits are spent only when the video is finished; if it fails, it costs nothing. Start with your photo.

• Genjutsu or Kling, if you already pay for one of them, want control over the trim, the aspect ratio and the number of takes, or want to put in a pet.
• A CapCut template, if you want something quick on your phone right now and do not mind a rougher swap.
• Burning Bridges AI Video, if you want a finished clip of one person in a few minutes without learning a tool or finding the source clip. See pricing for the packs.

If cost decides it, the free options guide compares what you get for nothing with what you pay for.