SeedRouter for Produce Audio Tracks

If you want to produce audio tracks from a prompt or work with existing audio material through a single API, SeedRouter gives you one place to call the models that support that job. Its audio API can generate full songs, voices, and dialogue from text, or separate stems from an existing track. When a model provides them, SeedRouter returns audio as tracks with duration, lyrics, and cover art.

That matters because the platform treats audio the same way it treats its image, video, and language models: one API key, usage-based billing, and a listed price before you send the request. Requests are charged when the result is delivered, failed requests cost nothing, and there is no subscription or monthly minimum. You can also track usage and cost per request in one dashboard, which is useful when audio generation is one part of a broader build rather than a separate tool you manage on its own.

For longer-running audio jobs, SeedRouter uses the same task pattern it documents for other asynchronous model calls. The request returns a task ID, and you can poll for status or receive a webhook when the output is ready. That makes it a fit for applications that can wait for completed audio rather than expecting an immediate inline response.

SeedRouter’s audio coverage is part of a broader model platform, not a standalone audio studio. The documented audio API is about model access, billing, and delivery; it does not add an editor or publishing layer of its own. If your product needs generated songs, speech, dialogue, or stem separation delivered through API calls, SeedRouter is the layer that connects those model outputs to your application.

Back to SeedRouter