An AI music studio.
Make it yours.
Start with a musical idea. Create an instrumental, hear a different direction and keep the versions you like. When you are ready for words, explore an AI singer or a saved voice you have permission to use.
Sign-in is required for private projects. Each new render has a credit cost shown before submission. Local generation uses AQUA's shared, capacity-limited queue.
Find the feeling.
Describe the mood, instruments and musical arc. Set the length, or start with a simple brief.
Hear another direction.
Make a new variation, compare complete takes and lock the version you want to keep.
Give the idea a voice.
Choose lyrics, language and an eligible singer for a new vocal arrangement. Earlier audio stays saved.
Text to music: start with a clear brief.
In a text-to-music workflow, a written musical brief guides a new render. Studio Lab keeps separate AI music variations in your project so you can compare complete takes and choose a favourite.
You do not need to write a technical production specification. Describe what the music is for, how it should feel and a few sounds you want to hear. Optional controls let you guide duration, tempo, mood, instruments, exclusions, genre, key, structure and arrangement.
“A soft cinematic instrumental. Begin with felt piano and solo cello, build gently with warm strings, then leave a quiet ending. No singing or drums.”This is an original example brief, not a guaranteed result. Musical settings guide a new generation rather than fixing every note or instrument.
Review the displayed credit cost, then submit. The project follows the real queue status while AQUA's local music backend renders the audio. A queued job is waiting for capacity; it has not started rendering yet.
Explore instrumental ideas or use the music prompt guide to make the brief more specific.
Compare AI music versions. Keep your favourite.
A variation starts from your selected version and a new direction, such as “make the ending bigger” or “try less percussion.” It creates a new complete render. The earlier version remains available, so you can compare the two and return to your preferred take.
- Rename takes, add notes and mark favourites.
- Compare complete versions with A/B playback.
- Lock the chosen instrumental and keep its place in the project's history.
- Return to the same account after a refresh to find your saved project and versions.
These are creative variations, not isolated edits to a recorded cello, drum or individual note. Listen to each result and check whether the requested change came through.
An AI singer, or a voice that is yours to use.
Choose an AI Singer or an eligible saved voice profile from your own account. A saved voice must be ready and authorised for synthetic singing. AQUA checks ownership, readiness and consent before accepting that voice for a new render.
Bring original lyrics you have permission to use, or develop lyrics with AQUA. Choose a supported singing language and describe the vocal direction. Your choices stay in the saved musical brief; review the actual sung words, pronunciation and performance before sharing the result.
Adding vocals creates a new complete arrangement. The original locked instrumental stays saved, but this workflow does not place a new singer over that exact backing recording. Studio Lab explains this and asks you to acknowledge it before generating the vocal version.
Read the saved-voice setup guide for reference recordings, permissions and supported voice use.
A considered finish. A real audio download.
Preview the completed song and use available WAV or MP3 downloads. Studio Lab's browser mix tools can reduce the whole track's level and apply a fade-in or fade-out, then create a WAV download of that mix.
Mix settings and export history are saved with the project. A browser-created mixed WAV downloads to your device; its audio is not uploaded back to AQUA or synced as a server-stored export. Keep a copy of the downloaded file.
The current tools adjust the complete mix. Separate singer and instrument levels, stem EQ, reverb, panning and combining sections from different takes are not available in this version.
Aqua SI, beside your creative work.
The project guide can help turn a vague musical aim into a clearer brief, explain settings or suggest a next variation. It uses AQUA's local 20B assistant through the existing queue.
The guide works with the project text and settings. It does not hear your audio, and a chat reply does not start a paid render. You choose which advice to apply and submit each generation yourself.
Learn about Aqua SI and its reasoning and document features.
Know what happens before you create.
Projects and saved voice profiles require authenticated account access. Completed songs begin in your private library. This public overview does not display customer projects or recordings.
A new variation or vocal arrangement is a separate generation with its own displayed credit cost. Credits are reserved when the request is accepted; eligible failed renders return the reservation. Account, storage, duration, daily allowance and queue limits still apply.
Generation can take minutes and longer while other work is waiting. Musical instructions are not a guarantee of an exact arrangement, perfect lyrics or a particular voice likeness. Check the finished result before relying on it or publishing it.
Use original or appropriately licensed words and music, and only voices you own or have permission to use. Read the voice-consent rules and privacy policy.