Find an AI audio workflow for your project
Audio models serve different purposes: creating music or sound, working with speech, transforming a recording or extracting useful information. Choose by the task and required input, then inspect the actual model controls.
Match the model to the audio task
Text to audio and music workflows begin with a written brief. Other workflows may require lyrics, a recording, a video or a reference track. Speech generation and transcription have different inputs and outputs, so review the workspace rather than treating every audio model as interchangeable.
Describe the sound clearly
For a creative audio brief, specify the mood, instruments or sound sources, rhythm and intended use. For example: A calm instrumental background with soft piano, light percussion and a gentle build for a product introduction. Confirm that the model supports the duration and output options you need.
Prepare source recordings
When a workflow takes a recording, use a file that meets its input requirements and contains the material you are entitled to process. Avoid unnecessary noise or unrelated sections where possible. Language, speaker and format support vary by model.
Listen before exporting
Review the credit estimate after login and keep the workspace open during generation. When an audio result arrives, use the player controls to listen through it. Check the beginning, ending and transitions before incorporating it into your project.
Explore available audio workspaces
Examples span the available model hubs. Compare the inputs and settings in each workspace to choose a workflow for your project.
Common questions
Can I generate music with AI?
Browse the available music and audio workspaces and choose a workflow whose inputs match your brief. Available capabilities and settings vary.
Can I upload my own recording?
Use a workspace that explicitly accepts source audio and follow its file requirements. Only upload recordings you are entitled to process.
Where do completed audio results appear?
The selected model workspace displays completed results and audio player controls. Saved requests remain available through your owned request history.