One face, then the mouth
Rap Video from a Photo
Make a rap video from a photo by uploading a picture you have permission to use, generating a track, and lip-syncing it.
A rap video from a photo is the Selfie mode. You upload one picture of a person who agreed, including you. The job writes a short music track from your lyrics, then lip-syncs that track to the photo. It is two steps because the character model refuses real faces, and the lip-sync model is a different one.
The photo needs a visible human face, decent light, and no crowd if you can help it. A poster, a logo, or a shot from a concert will not become a clean performance. Duo is the same pipeline with two photos, not a slideshow of ten.
Permission is the whole rule. A photo you took of a friend is allowed when they said yes. A celebrity still, a classmate's profile picture, or a picture of a minor you do not have a parent's OK to animate is not. The form asks you to confirm that before the job starts.
Your photo is stored so the render can read it. It is not a public gallery. The mock mode on a local machine plays a labeled sample instead of your face, so you can learn the flow before any paid model is connected.
01
Write the verse first
The mouth will be busy for twelve seconds. Give it lines that fit.
02
Upload one permitted photo
JPEG, PNG, or WebP. Confirm you can use the face.
03
Wait for track, then lip-sync
The status moves from music to lip-sync. Failed jobs return the credits.
Questions
Why is a photo two steps?
The face model and the character model are different. Selfie makes audio, then animates the photo to it.
Can I use several photos as scenes?
No. Selfie is one photo. Duo is two. This is not a multi-scene editor.