The AI rap duo trend, starring the two of you

How to make a Rap Duo video yourself

You do not need this page to make one. This is how the video is put together, with the prompts for its one shot, so you can build it with any picture model and image-to-video model you like.

1. Step 1

   Draw one still from both photos

   Use a picture model that takes reference photos and send both at once, naming the first as image 1 on the left and the second as image 2 on the right. Ask for the two of you from the waist up, bodies running out of the bottom of the frame, faces to the lens, lips slightly parted, a microphone hanging between you level with your chins, and one solid colour behind.

2. Step 2

   Write the verse into the motion prompt

   Give the still to an image-to-video model that makes sound with the picture. Put the four bars in quotes, say who raps which two, name the voice and the beat, and ask for lips and jaw moving with every syllable. Swap "female" for "male" in both places for a male voice. Keep it to about twelve seconds.

3. Step 3

   Keep it to one take and post it with its own sound

   Use the clip as it comes back: one shot, no cuts. A second clip would arrive with a different beat and different voices, so the verse would change halfway through. Do not lay a song over it either; the rap is the sound, and it only matches the mouths because it was made with them.

The shot, second by second

The prompts

The whole video is one shot, made in two steps. Give its still prompt to a picture model with the photos listed with it as references, then give the picture it draws to an image-to-video model with the motion prompt. Add these endings to both prompts: they set the look and keep the faces the same from the first frame to the last.

1. 0–12sTrading bars

   Photos to attach, in orderImage 1: Who raps firstImage 2: Who raps second

   Still prompt

   Copy

   A medium shot of two people standing upright side by side, close together, their shoulders almost touching, seen from the waist up and filling the frame: their heads and faces in the upper half of the frame, their bodies running down out of the bottom of the frame, both faces large and sharp. On the left of the picture is the person in image 1, in their own clothes from image 1; on the right of the picture is the person in image 2, in their own clothes from image 2. Both stand facing the camera in a three-quarter view, their faces turned toward the lens so each face is seen from the front, mouths in full view, their hands down by their sides. Hanging in the air between their two faces, level with their chins and covering neither face, is a single plain black studio microphone with no logo, badge or writing on it, its grille pointing down at the floor, held only by its thin black cable, which runs from the top of the microphone straight up and out of the top of the frame. Nothing is below the microphone: no stand and no cable underneath it. Both have serious, focused expressions, lips slightly parted, as if about to rap. Behind them is a bare studio painted one solid, saturated deep cobalt blue, a seamless cobalt-blue wall, evenly lit, with no furniture, no equipment and no writing. The blue wall is all there is behind them, edge to edge, with no lights, lamps, ceiling or rig anywhere in the picture. Nobody else is in the picture.

   Motion prompt

   Copy

   The two of them trade bars into the hanging microphone, rapping a verse in a confident female voice with a tight, rhythmic rap flow, over boom-bap drums and a deep 808 bass. The one on the left raps the first two lines, their lips and jaw moving in sync with every syllable, while the one on the right nods to the beat with their mouth closed: "Two of us, one mic, and the red light is on, / blue walls glowing and we rocking till the dawn." Then the one on the right takes over and raps the last two lines, also in a female voice, their lips and jaw moving in sync with every syllable, while the one on the left nods to the beat with their mouth closed: "You bring the rhythm and I bring the rhyme, / side by side and we land it every single time." Both heads nod on the beat the whole time. Neither of them smiles or laughs at any point: serious, focused expressions throughout. The camera holds still on both of them as framed in the first frame, both faces in view the whole time, with only a slight handheld sway: no zoom, no push-in.

Where it goes wrong

• A rap made separately will not match the mouths

  Generating the rap on its own and laying it over a silent clip looks wrong within a second: the mouths move, but never with the words. Write the verse, the voice and the beat into the motion prompt and let one model make the picture and the sound together.

• Tell the one waiting to keep their mouth shut

  Say only who raps which lines and the model often moves both mouths to one voice. Say that whoever is not rapping nods to the beat with their mouth closed, in both halves, and the mouths switch over where the words do.

• Hold the camera, and hang the microphone the right way

  Left to itself the camera can push in on one face and lose the other for most of the verse: ask for both faces in view the whole time, no zoom. Do not write that the microphone hangs upside down either, or the whole picture can come back upside down; say it hangs on a cable from above.

• Only people who said yes

  Use your own photos or photos of people who agreed. Do not use celebrities, real rappers, the performers the trend first copied, or anyone who has not agreed to be in the video: a realistic video of a real person rapping words they never said can hurt them, and platforms take it down.

• Label it as AI-generated

  Switch on TikTok's AI-generated content label before you post. A realistic clip of two people rapping a verse they never recorded is exactly the kind people ask about, and viewers trust a post that says so.