The AI rap trend is the fast-climbing name for a whole category of AI videos: clips that drop anyone — you, a friend, a celebrity, a game character — into an iconic rap performance. Nothing is filmed. A reference clip and a character image go into an AI video tool, and the model re-stages the scene around your chosen cast. The format dominating the wave right now is the Hotel Lobby restyle, which is why the same meme also travels as the "Migos AI video" — different search words, identical trend.
New here? The homepage overview sums up the Hotel Lobby wave on one screen, and the AI video trends hub tracks every new meme as it breaks — including the car driving AI video that broke alongside this one.
Think of it as a format family rather than one video. Every member follows the same recipe: take a famous rap performance, run it through a video-to-video model, and redraw who is in frame while the original staging, motion and camera work survive intact. The result reads as a real music-video moment, except the performers are GTA characters, anime figures — or you.
That is what separates the AI rap trend from ordinary AI clips: the source material is shared, so every output is instantly comparable against the template and against other people's casts. Comparison drives comments, comments drive reach. The Hotel Lobby COLORS restyle is currently the biggest cluster inside the category, and tutorials for it are the ones spreading widest.
Every trend needs a template, and the COLORS x Rap Caviar performance of Hotel Lobby by Quavo & Takeoff has the anatomy of a perfect one: a small cast, strong lighting, stable camera motion and instantly recognizable staging. Video-to-video models keep exactly those qualities when they restyle a scene, so outputs look deliberate instead of mushy — and because everyone restyles the same clip, the wave compounds.
The naming splintered as it spread: migos ai, migos ai video, the Migos AI trend, hotel lobby ai — and now the umbrella phrase ai rap trend, which is the category-level way to search the whole movement. Same meme, same workflow, different search vocabulary.
The mainstream route runs through Higgsfield's Genjutsu video-to-video model. The full walkthrough, including the mistakes that waste your first renders, lives in the Higgsfield Hotel Lobby tutorial; the short version:
Want to see the model in motion before spending credits? Watch Higgsfield Genjutsu in 7 Minutes. Written walkthroughs of the same flow are circulating from Nate Herkelman on LinkedIn and Kienobi on Substack, and short-form tutorials keep spreading across TikTok and YouTube.
Genjutsu is preset-driven — you can generate with no prompt at all — but a short description of the restyle usually sharpens the result. A few habits separate good casts from mushy ones:
Yes. A new Higgsfield account includes free trial credits, and that is enough to render your first clips — the mainstream workflow produces a postable video before any payment comes up. If you want routes that stay free past the first renders, the free alternatives comparison lines them up.
No. There is no timeline, no cuts and no effects stack — the model produces the finished scene from your reference clip and character image. Your work is choosing the reference, picking the cast, and iterating until the render looks right.
Where the trend lives: TikTok and Instagram Reels. The tutorial ecosystem around it — LinkedIn write-ups, TikTok walkthroughs, YouTube guides — feeds discovery, but the audience is on the short-form platforms. Post while the wave is climbing; early posts ride the search wave, late posts compete with thousands of clones. Track what breaks next at /trends.
Be careful here. Generating the video with an official tool is one thing; the music you attach is a separate right. Do not publish with the original track for commercial purposes — pair the clip with royalty-free or platform-library audio instead, and follow each platform's music rules. The visual trend travels fine on trending free audio, which is what most viral posts use anyway.