umayma

An AI studio for making songs and cloning voices. Enter the password to come in.

You're in. What should we call you? Your name goes on the songs you make.

umayma
Music Voice

What do you want to make?

Make a song Describe it, pick a sound, and get a full song with vocals and cover art in a few minutes. Clone your voice Record 30 seconds, then make your voice say anything, in any language.

Made in the studio

Which song should we restyle?

Sound
Mood
Voice
Lyrics language

Writes the title, sound and lyrics from your idea in about 20 seconds. Edit anything after.

This is what the music model hears. Leave it empty and your picks above are used.

Leave it empty and the AI writes the lyrics for you.

Takes
More settings
Length

Shorter songs are ready sooner.

Model
File

Songs

    No songs here yet. Describe one and press Make song.

    Clone your voice.

    Read this out loud

    0:00Tap to record. Aim for 30 seconds.

    One speaker, no music behind it. Clearer audio makes a closer clone.

    Your voice is ready.

    Listen to it say hello, then make it say anything.

    Your voices

      Voices you clone or describe show up here. Only you can see and use them.

      Make it talk.

      Voice

      Clone your own voice to hear yourself. These studio voices work right away.

      Say it in

      Pick a language and your text is translated first, then spoken in the same voice.

      Feeling
      Speed
      Effect

      Your clips

        0:00/0:00

        0:000:00
        Download
        Sound description

        Settings

        Studio GPU

        Checking the studio GPU…

        MiniMax Music 3 on Modal H100s, about $4 an hour for each one that's on. Up to 3 GPUs, 4 songs at a time on each. Each one turns itself off after 3 idle minutes.

        Music engine

        MiniMax key

        Checking the MiniMax key…