AI rap video generator: you, rapping, with your real face

Most AI rap video apps animate one photo, so the face is only as good as that photo and drifts as soon as the head turns. RefVid works from a saved set of 1 to 5 photos per person and sends them as references with every clip. The model sees the face from several angles, so it holds through head nods, hand moves and a full 15-second take.

Because each person is saved once, the next clip starts from the same references. A weekly rap series, a birthday verse and a reply to a comment all show the same two people.

The AI rap video that is everywhere right now puts two people side by side in a bright orange studio booth, each at a microphone, trading lines to the camera. It is a duet, so the tool has to keep two different faces apart in one shot. RefVid takes up to three saved subjects per clip and keeps each one tied to its own photos.

The rap duo template has the booth, the microphones, the camera and two lines already written. Save yourself and your partner, open the template, change the bars if you want, and generate. Prefer a different look? Write a rooftop at night, a school hallway or a hotel lobby instead of the orange walls.

Three to five clear photos per person: a straight-on face, a three-quarter view each side, and one waist-up shot. Even light, no sunglasses, nothing covering the mouth. The mouth matters more here than in most videos, because rapping is all mouth movement.

Pets, mascots and toys work the same way: save them as a subject from a few photos and write what they do on the beat. A dog that nods along while you rap is one of the easiest versions to get right.

Put each line in quotes and name who raps it. The model raps quoted lines as written. Keep each line to one bar, about 8 to 12 syllables, with a clear rhyme at the end. English comes out clearest. Two or three short lines per person fit in 15 seconds; a full verse does not.

A prompt that works: "{you} and {friend} stand side by side in an orange studio booth and rap over a hard trap beat, nodding to the rhythm. {you} raps: 'Same face every frame, yeah we never switch it up.' Then {friend} raps: 'Two photos in and now the whole world's watching us.' Static vertical medium shot." Describe the beat (trap, boom bap, drill) and the energy (laid back, hyped) to steer the sound.

A clip is up to 15 seconds, vertical 9:16 for TikTok, Reels and Shorts. The beat and the vocals are generated new with the video, so the clip carries no copyrighted song and is safe to post as it is. If you want a trending sound instead, mute the clip and add the sound inside TikTok or Instagram when you post.

For a longer song, make several clips with the same two people and join them in any editor; the faces match across clips because they come from the same saved photos.

New accounts get free credits for a first short clip, and RefVid never adds a watermark. A full 15-second rap video can be bought as a single video with no plan, or paid from a credit pack. Current prices are on the pricing page.

Use photos of yourself, or of friends who agreed to be in the video; the studio asks you to confirm this for every adult. Don't make anyone rap words they would not say. TikTok, YouTube and Instagram ask you to label realistic AI video, so switch on the AI label when you post.