Music and sound effects

Open Create → Audio to make music or sound effects when a compatible model is published. Audio generation is on by default; administrators can publish a model after configuring its channel. Your model determines the available controls and the price.

Create audioCopy Anchor Link

1. Choose Music or Sound effects, then select a model.
2. Describe the sound. For music, include mood, instruments, genre and use. For effects, describe the source, surroundings and changes over time.
3. Adjust the supported settings. Some effects accept durations below one second and a loop option. Models with prompt-guided duration choose the final length themselves.
4. Review the credit quote, then select Generate audio. Preparing a template and requesting a quote do not charge credits.
5. Listen to the saved result. Download the original file or find it later in your project's library.

For a song, use your original lyrics in the separate Lyrics field when the model supports it. Instrumental mode creates music without sung lyrics. A loop setting asks the model for a smooth repeat; listen to the transition before using it in a finished project.

Templates and recoveryCopy Anchor Link

The Audio templates include product music, podcast intros, game music, notification sounds, environmental ambience and songs from your lyrics. Fill the template, select a compatible model and prepare it. The next step opens Create so you can review the current settings and quote before generating.

Accepted tasks continue after you leave the page. If a submission is interrupted before its response arrives, use Check the original submission. Artgen keeps the original request key to avoid charging for a second task. Refreshing the page reads saved jobs and does not generate new audio.

Cancellation is available only before the task is dispatched. Failed tasks return the applicable reserved credits. If a provider cannot confirm a submission, Artgen can refund the platform reservation; that is not a guarantee the provider cancelled its own work.

Outputs and limitsCopy Anchor Link

Audio files are private to your account. The library shows measured audio length, which may differ from a requested length. Available formats depend on the model; the initial integrations use MP3 and, for supported models, WAV. This release creates standalone audio. Audio uploads, reference audio, video synchronization, stem editing and mixing are not available yet.

Start with music or sound effects.