The carpool karaoke AI trend is a short AI video of friends or pets rapping along in an SUV, styled after Migos' Carpool Karaoke segment from 2018. On videosi you make an 11-second silent version from two photos, or one photo of you both, and add Walk It Talk It yourself where you post.
This guide goes past the basics on the trend page. It covers where the format comes from, how the clip is built, how the seats and photos work, and what to check before you post.
Key takeaways
- The trend copies a daytime SUV singalong with dashboard cameras, based on Migos' Carpool Karaoke of 13 November 2018.
- videosi needs two photos, or one photo of you both: you in the passenger seat, a friend or pet driving.
- The 11-second clip is silent and not lip synced, so you add Walk It Talk It from the platform's sound library.
- No show footage, no Migos or James Corden faces and no celebrity photos are used or accepted.
- A standard 11-second video is 10 credits, with packs from $6.99.
In this guide
- Where does the carpool karaoke AI trend come from?
- What is the song, and who is actually on it?
- How is the videosi version built, and what do the 11 seconds contain?
- Is it lip synced to Walk It Talk It?
- Which two people, or pets, should I put in the front seats?
- What if I only have one photo of me and a friend?
- Why only two people when app templates take four?
- How does it differ from the White Chicks AI car video?
- Can I use celebrity photos, or the Migos themselves?
- What should I check before posting?
- What does it cost, and is there a free version?
- Sources
Where does the carpool karaoke AI trend come from?
It comes from one TV segment. On 13 November 2018, Migos (Quavo, Offset and Takeoff) rode with James Corden on Carpool Karaoke, and The Fader reported that they rapped along to "Bad and Boujee" and "Walk It Talk It". The clip also included a quiz on dabbing and a bit about counting handfuls of cash.
Eight years later, the segment gave AI creators a format that is easy to copy: a car, a few friends, small cameras cutting between seats, and everyone throwing hand gestures to the beat. The earliest AI versions and tutorials we found on TikTok are dated 23 and 24 September 2026, and we could not identify one original post or creator, so we do not credit anyone for starting it.
What is the song, and who is actually on it?
The song is Walk It Talk It, by Migos featuring Drake. According to Wikipedia, it is the sixth track on Culture II, it came out as a single on 18 March 2018, and it peaked at number 10 on the Billboard Hot 100. The Carpool Karaoke segment is where many people first saw the group perform it in a car.
The title is also why you see two search phrasings: "walk it talk it" and "walk it like I talk it". People use both for the same trend. If you search for the sound inside TikTok, Reels or Shorts, try both spellings and pick the version that has the most clips under it.
How is the videosi version built, and what do the 11 seconds contain?
The carpool karaoke AI video on videosi is a new daytime SUV scene, not a copy of the show. It runs 11 seconds and cuts between a wide dashboard shot of both front seats and a close-up of each person, so the rhythm feels like the segment: a passenger verse, a driver line, a moment where both bounce their shoulders, and a last line that ends in a grin.
Two characters sit in the front seats, the passenger on the left of frame and the driver on the right with one hand on the wheel. The back seat stays empty and the car keeps driving calmly. There is no text, no lyrics on screen and no show branding, so the clip reads as your own video rather than a fan edit of the original.
Is it lip synced to Walk It Talk It?
No. The file is a silent MP4. The mouths move like rapping and singing, but they are not matched to the words of the song, and the model never hears the track. That is a deliberate choice, because it keeps the clip free of any audio that belongs to someone else.
In practice it matters less than you would expect. The trend runs on energy: head bobs, pointing and hype gestures. When you add the song in your posting app, start it on the hook so the first beat lands on the opening wide shot. A short, loud section of the track hides small differences between mouth shapes and lyrics better than a quiet verse does.
Which two people, or pets, should I put in the front seats?
Photo 1 is the passenger, the one who raps down the lens in the close-ups. Photo 2 is the driver, who sings with eyebrows up and keeps eyes on the road between lines. Give the passenger seat to whoever has the bigger stage personality, since that person gets the longest close-up at about 7 to 9.5 seconds.
Either seat can be a person or a pet. A dog driving while you rap beside it works well, and two pets in the front seats also works. The pet sits upright with a seatbelt, bobs its head and keeps its breed and markings. Choose a front-facing photo of the animal with the face clearly lit.
What if I only have one photo of me and a friend?
Switch to One photo together and upload a single picture of both of you. The passenger is the person on the left of the photo and the driver is the person on the right. If your photo has them the other way round, flip it before uploading.
Keep the photo to two faces, similar in size, both looking toward the camera. A crowd shot is a poor fit, because the scene has exactly two characters. Two separate photos usually give more control over each face, so use the shared mode when it is the only picture you have.
Why only two people when app templates take four?
Some app templates swap up to four faces into a car, including the back seat, which mirrors how many people rode in the original segment. videosi makes the front-row version only, and the back seat stays empty. The reason is reliability: every extra face in an 11-second clip is another chance for faces to blend or swap, and two characters keep both likenesses stable through close-ups and side angles.
If you want a bigger crew, make two clips with different pairs and cut them together in your editor. Because every clip uses the same SUV and the same camera language, they join up naturally.
How does it differ from the White Chicks AI car video?
They are both car singalongs, but they are different scenes. The White Chicks AI video is a night convertible where one person sings A Thousand Miles and the passenger gives a side-eye. The carpool karaoke version is a bright daytime SUV where both people perform, with dashboard cameras cutting between them.
Choose White Chicks when you want a comedy reaction, one big singer and one deadpan face. Choose carpool karaoke when you want two people hyping each other up. Quavo and Takeoff also have another AI meme on videosi, the hotel lobby AI video, if you want a second clip in the same rap-meme family.
Can I use celebrity photos, or the Migos themselves?
No. videosi does not accept photos of celebrities, public figures or minors, and the page never uses Migos' or James Corden's faces. Many viral posts put athletes or other famous people in the car, but a realistic video of a real public figure is exactly the kind of clip that misleads people.
Use adults who agreed to it, or your pets. videosi is also not affiliated with Carpool Karaoke, The Late Late Show, CBS, Apple or Migos. Those names only describe the trend, and the generated scene uses no frames, faces or audio from the show.
What should I check before posting?
Run through four things. First, add Walk It Talk It from the platform's own sound library, not as a file inside your edit, so the platform handles the licensing for that use. Second, switch on the AI-generated label in TikTok, Reels or Shorts. Third, check the faces in the side angles, where likeness drifts most. Fourth, trim the clip so the song starts on the hook.
If a face looks wrong, the usual causes are sunglasses, a tight crop, a dark photo or a side-on pose. Retake the photo in even light with head and shoulders in frame. Likeness is not guaranteed in 11 seconds, and a failed, timed-out or blocked render returns its credits.
What does it cost, and is there a free version?
There is no free generation and no free trial. A standard 11-second video (Standard model, 720p) costs 10 credits. Credit packs start at $6.99 for 10 credits and never expire, and plans start at $29.90 a month. Better models and higher resolutions cost more credits per video.
Because most people need a couple of tries to get both faces right, buying a small pack and testing with clear photos is cheaper than rerunning with poor ones. When you are ready, the carpool karaoke AI page takes two photos and returns the clip in a few minutes with no prompt to write.
Sources
Background on the 2018 episode comes from The Fader's report on the Migos Carpool Karaoke appearance, and song details come from Wikipedia's page for Walk It Talk It. To make your own, open the carpool karaoke AI page, add two photos and press Generate.
