Skip to main content

Video trend

Bandit Lip Sync AI Video: How the Mouths Follow the Song

A lip-sync trend lives or dies on the mouths. In the Bandit duo clip the lead man mouths a verse straight to the camera, the sidekick raps one line in the foreground near the end, and two tight close-ups put the face right in the frame. This page explains how the mouths in your version are produced, which shots matter most and what to look at before you post it.

Final video
21 seconds
Your input
2 photos
Quality
720p
THE ORIGINAL21 SEC / 16 / 9
Bandit Rap DuoYour photo. This scene.

How the lip sync works and what to check

01

What the model follows

The video model receives the reference clip and your two photos, with no audio or voice sample. It follows the mouth movement it can see in the reference clip, puts your people in those mouths, and Oakgen puts the original song back on the finished file. Because the model never hears the song, lip sync is not guaranteed to match every syllable.

02

Where the lead's mouth matters

The Lead rapper lip-syncs to camera in the glass-block studio from about 3.3 to 5.5 seconds, in the tight close-up against blue glass blocks from about 13.5 to 16.1 seconds, and in the final close-up, pointing at the camera, from about 19.9 to 21 seconds. In between he gestures, points to his head and walks the lawn.

03

Where the sidekick's mouth matters

The Sidekick dances behind for most of the clip and raps in the foreground of the second studio scene, from about 18.3 to 19.9 seconds, with the lead behind. It is a short moment, so check it closely: it is the only time the second person's mouth is the focus.

04

Photos that give the mouth the best chance

Use a sharp head-and-shoulders photo, face forward, hair fully visible, mouth uncovered, with no hat and no sunglasses. Keep hands, microphones and drinks away from the mouth. A relaxed, slightly open or neutral expression works better than a wide smile or a pout, because the model has to reshape the mouth for rapping.

05

How to check it

Watch it twice: once with the sound off, looking only at the mouths in the three close shots, then once with the song on. The mouths should open and close on the beat, not drift into a smile. If a mouth looks wrong in one shot but fine in the rest, that is a sign to generate again with a clearer, more front-facing photo.

06

If it is slightly off

Trim the weak shot in your editor and cut away to a wide shot, such as the night lawn or the ATV wheelie close-up, which show no mouths. The template has 10 shots, so there is room to tighten without losing the structure.

07

Who is rapping

Neither of the people in the reference clip is real footage of anyone, and the voice you hear is the song's own recording, kept as it was. It is not a voice Oakgen made or cloned, and it is not your voice or your friend's. Label the post as an AI parody so nobody mistakes it for the artists.

08

Make it

Open the Bandit Rap Duo trend, add the Lead rapper's photo with the mouth clearly visible, then the Sidekick's, and generate. Check the three mouth-heavy shots before you share.

Explore the Bandit Rap Duo trend

How to make it

Add a photo for each role shown below. All roles need a photo.

The roles in the Bandit Rap Duo original reference
Lead rapper in the Bandit Rap Duo source video
Lead rapper
Sidekick in the Bandit Rap Duo source video
Sidekick
  1. 01

    Add the two photos

    Use a head-and-shoulders photo each for the Lead rapper and the Sidekick, face forward, hair fully visible and no hat. Use two different photos so the duo reads as two people. Use your own photos or ones of people who agreed to appear.

  2. 02

    Choose the quality

    Pick 720p or 1080p and review the quote. It covers the full 21-second video with both people replaced.

  3. 03

    Check the mouths and post

    Watch the studio lip-sync and the two close-ups with sound, then download your landscape 16:9 MP4 with the original soundtrack and label it as an AI parody when you post.

01

Your faces, not theirs

The template follows the parody clip's shots and moves, but both rappers come from your photos: face, hair, build and outfit in every shot. The reference men's suits, shirts, ties, glasses, ring and trainers are not kept, and no one's likeness is copied from the reference.

02

The song stays

Your video keeps the reference clip's original soundtrack, a verse of the song Bandit. No separate audio or voice sample goes to the video model; the mouths follow the rapping in the reference clip and the original track is put back at the end. Oakgen does not clone or generate a voice.

03

Ten shots in 21 seconds

A night-lawn approach, a glass-block studio lip-sync, a rider passing, a wheelie close-up, a dry-grass backdrop, a grey wall, a tight close-up on blue glass, a cuff-and-turn on the lawn, the sidekick's rapped line and a final point at camera.

04

A parody, labelled as one

The preview is an AI-made parody clip, not real footage of any real person, and none of the song's artists made or endorsed it. Backdrops, jackets and props are asked for with no readable lettering or logos in your version.

Photo tips

  • One person per photo
  • Two different photos
  • Face forward, hair visible, no hat
  • Keep hands and drinks away from the mouth
  • Your own outfits stay in every shot
  • Label it as an AI parody

Questions & answers

What is the Bandit Rap Duo AI trend?

An AI-made parody in which two men in suits lip-sync and dance to a verse of the song Bandit by Juice WRLD and YoungBoy Never Broke Again on a night lawn, in a glass-block studio, in front of a dry-grass backdrop and against a grey wall, while riders pop wheelies on ATVs behind them. Oakgen's template keeps the shots, the cuts and the soundtrack and puts the people from YOUR two photos in the lead rapper's and the sidekick's places.

Is the preview real footage?

No. The preview is an AI-made parody clip, not real footage of any real person, and none of the song's artists made or endorsed it. Oakgen is not affiliated with or endorsed by the song's artists, their labels, or any other artist, label, brand or sponsor, and your video does not show anyone from the reference: the two rappers are the people in your photos.

What song is it?

Bandit by Juice WRLD and YoungBoy Never Broke Again, a single released on 4 October 2019 and produced by Nick Mira. The clip lip-syncs to the song's second verse. It is an explicit track, so we do not quote it.

Does my video keep the voice?

Yes. Your video keeps the reference clip's original soundtrack, the song. Your photos change how the two rappers look, not the words or the voice, and the template does not take a new script or voice recording.

Is the voice sent to the AI model?

No. No separate audio or voice sample is sent to the video model, and whatever sound the model returns is thrown away. The mouths follow the rapping in the reference clip, and the original soundtrack is put back on the finished video. Oakgen does not clone or generate a voice for this template.

Are there subtitles?

No. This reference clip has no on-screen subtitles, and the template asks the model for no captions or text. The song carries the words, and you can add your own captions in your editor before you post.

Do I need two photos?

Yes. The Lead rapper and the Sidekick each need a photo, because the template replaces both men in the reference clip completely, so neither of them is kept in your video. Two different photos make the duo read as two people.

Who should be the lead rapper?

The person with the more expressive face. The lead gets the lip-sync to camera, the pointing and the two tight close-ups. The sidekick dances behind, stands in the dry-grass shot and raps one line in the foreground near the end, which makes it the funnier role for a friend.

Do we have to wear suits?

No. There is no costume. Each person wears the outfit from their own photo in every shot, with their own hair and accessories, instead of the reference men's suits, ties, glasses and ring. A photo that shows only the upper body gets plain dark trousers and shoes.

Who else is in the video?

The ATV riders and any other background people stay as they are in the AI parody. Only the two rappers are replaced, and nobody else takes your face.

Are brands or logos in my video?

None is meant to be. The template asks for the jackets, the backdrops and every prop to show no readable lettering, logos or numbers. Check your result for stray marks before you post. The trend is not connected to any artist, label, brand or sponsor.

Can I put a pet in it?

Yes. A clear photo of a dog or cat can take either role, with its own face, fur and markings. Check the mouth in the lip-sync shots and the paws in the pointing shots before you post.

Whose photos can I use?

Your own, or photos of people who agreed to be in the video. Do not upload photos of artists, other public figures or anyone who has not consented, and label the video as an AI parody when you share it.

What kind of photo works best?

A sharp head-and-shoulders photo, face forward, hair fully visible, mouth uncovered and no hat or sunglasses. One person per photo, with no hand or drink near the mouth.

What quality can I choose?

720p or 1080p. 720p is enough for a phone feed; 1080p keeps the faces sharper in the two tight close-ups.

How long is the video?

21 seconds in landscape 16:9, with 10 shots. Place it on a vertical canvas for Reels, TikTok and Shorts.

Is the preview generated by Oakgen?

No. It is the AI-made parody reference clip the trend comes from, used for the shots and timing, not an Oakgen result.

How much does it cost?

1,115 credits per 720p video. The quote covers the complete 21-second video with both people replaced. Check it before generating.

Ideas for your next video

Plan a story, keep your character recognizable, and make the finished clip your own.

Explore the video trend playbook →

Make your Bandit Rap Duo video

Upload 2 photos. Everything else is set.

Make this video