An AI rap video turns a photo or two into a short music video: two people (or one) on a stage, rapping a song that was written about them. People use it for birthday messages, roasts, anniversaries, group chats and even small-business promos.
This guide walks through making one with RapVideo AI, from the photos to the download, plus what it costs and what to do when a result looks off.
What you need
- One or two photos with clear faces
- A browser on your phone or computer. There is no app and nothing to film.
- A topic idea: a name and one funny or sweet detail is enough
- Credits for the render (see the cost section)
How it works, in one paragraph
In the New song style, a language model first writes rap lyrics from your topic. Then a video model takes your photos and the lyrics and generates the picture, the two voices and the beat in one pass, so the lip movement and the music are made together. The voices are AI voices; there is no voice cloning, so it will not sound like you or like any real artist.
The generator also has an Original remake style that puts your faces into the original 14-second "Hotel Lobby" clip with the real song. This guide is mostly about New song, because that is the one that makes a rap about you. If you are here for the trend itself, see the Hotel Lobby guide.
Step 1: Open the generator and add photos
Go to the generator on the home page. You will see three ways to add people:
| Option | What to upload | Good for |
|---|---|---|
| Two photos | One person in each photo. Photo 1 goes on the left, photo 2 on the right. | Friends or a couple who don't have a good photo together |
| One photo of both | One photo with both faces clearly visible | A couple or best friends with a favorite picture |
| Solo | One photo of one person | Birthday shout-outs, a personal hype song, a business owner |
The two photos don't need to be taken together. A selfie from today and a friend's photo from years ago work fine.
Photos that work best: one clear, front-facing face per photo, even light, no sunglasses or masks, no heavy beauty filters. Half-body shots look best on stage. Photos must be under 15 MB. Only use photos of people who agreed to be in the video.
Step 2: Pick a stage
In New song you choose where the performance happens:
- Hotel Lobby (orange): the orange booth with one hanging mic, the look from the viral trend. This is the default.
- Luxury Lobby: a fancier lobby.
- Studio Booth: a recording-studio look.
Original remake keeps the original set, so there is no stage to pick.
Hotel Lobby is the safe pick if you want something people recognize. Studio Booth suits a "we made a song" feel, and Luxury Lobby works well for birthdays and anniversaries.
Step 3: Pick quality and length (optional)
The default is a standard video: Standard model, 720p, 12 seconds, 9:16 vertical, 10 credits. Most people never change it. If you want to, open the quality panel and you can choose:
- Model: Standard, Fast, Pro or Ultra
- Resolution: 480p or 720p on every model, plus 1080p on Pro and Ultra
- Length: 5, 8, 10, 12 or 15 seconds, plus 20 or 30 seconds on Ultra
- Format: 9:16 (vertical), 1:1 (square) or 16:9 (wide). Every format costs the same.
The price on the panel updates as you change options, and that is the price you pay. Longer videos, higher resolutions and bigger models cost more.
Original remake always renders the fixed 14-second clip, so you can't change its length.
Step 4: Write a topic
This step is for New song. Original remake raps the original lyrics, so it has no topic box.
This is where the video becomes yours. Type a topic of up to 120 characters. The lyrics are written around it, so real names and one specific detail make the biggest difference. You can also tap a ready-made chip (Birthday, Anniversary, Best friends, My business) and edit it.
The rap follows the language of your topic: write it in Spanish and you get a Spanish rap.
Ideas by occasion:
| Occasion | Example topic |
|---|---|
| Birthday | "Jake's 30th birthday, the king of late replies" |
| Roast | "Maya says she's 5 minutes away, she hasn't left the house" |
| Couple | "Our 5-year anniversary, still fighting over the remote" |
| Best friends | "Best friends since 2nd grade, still the loudest in the room" |
| Business | "Our coffee shop on Main Street has the best iced latte in town" |
| Coworkers | "Friday deploy survivors, the build is green" |
Keep it short and concrete. "Happy birthday" gives the model nothing to work with; "Happy birthday Lena, who still owes me twenty dollars" does. You can also leave it empty for a general hype song.
Step 5: Generate
Press the button. A render takes about 2 to 3 minutes. You can leave the page; the video will be waiting in My videos when it is done.
Step 6: Watch and download
Play it back, then download it. Paid videos download without a watermark, ready to post or send in a group chat. If you want a different version, change the topic or the stage and make another one.
How much it costs
| Video | Credits |
|---|---|
| Standard, 720p, 12 s (the default) | 10 |
| Standard, 480p, 12 s | 5 |
| Fast, 720p, 12 s | 17 |
| Pro, 720p, 12 s | 20 |
| Pro, 1080p, 12 s | 50 |
| Ultra, 720p, 12 s | 37 |
To buy credits:
- The single pack is $9.99 and covers one standard video.
- The Basic plan is $29.90 a month for 100 credits, about $2.99 per standard song.
- Pack credits never expire. Plan credits refresh every month.
- Failed renders are refunded automatically, so a render that breaks never costs you.
See pricing for every pack and plan.
Troubleshooting
The faces look blurry or not quite right
Almost always the photo. Use a sharp, front-facing photo in good light, with the face not too small in the frame. Avoid sunglasses, hats that shade the eyes, masks and strong beauty filters. Tiny faces cropped from group photos lose detail.
One person looks wrong in a two-person video
Check that each photo has exactly one person and that both faces are clear. If you use "One photo of both", make sure nobody's face is covered or turned away.
The lip sync looks off
The voices and the picture are generated together, so the mouth usually follows the words. If it drifts, a clearer, front-facing photo helps most. You can also try a different model in the quality panel.
The render failed
You were not charged: the credits go back to your account automatically. Try again, ideally with a different photo.
My topic was rejected
Some topics can't be used. Nothing is charged; rewrite the topic and try again.
The rap is in the wrong language
The rap follows the language of your topic. Write the topic in the language you want to hear.
FAQ
Can I make an AI rap video of me and my friend?
Yes. Use "Two photos" with one of you in each, or "One photo of both" if you have a picture together. The photos don't need to be taken together.
Can I make one by myself?
Yes, choose Solo and upload one photo.
Will it sound like me?
No. It uses AI voices, and there is no voice cloning.
Can the rap be about anything?
Mostly, yes: birthdays, roasts, couples, friends, coworkers, your business. Keep it kind to the people in it, and get their consent before you post.
How long is the video?
A standard video is 12 seconds. You can pick from 5 to 15 seconds on most models, and up to 30 seconds on Ultra.
Do I need to download an app?
No. RapVideo AI runs in your browser on a phone or a computer.
Is it free?
No. A standard video costs 10 credits, and failed renders are refunded automatically.
Is this the same as the Hotel Lobby trend?
Hotel Lobby is one of the stages. If you want the trend specifically, read the Hotel Lobby guide or see the Hotel Lobby AI page.
Make yours
Pick a photo or two, think of one detail about the people in it, and try it on the generator. For more ideas, see the rap duo page, the Migos AI page and the Hotel Lobby AI page.
RapVideo AI is independent and not affiliated with Migos, Quavo, Takeoff, Offset, COLORS or any song's rights holders.