API Overview

The Mossland API uses /v1 audio endpoints. You call the endpoint that matches the capability you need, select a model with model, and receive results via a synchronous response, an asynchronous task, or an explicit status query.

Core concepts

• Endpoint – The interface entry point you request, for example POST /v1/audio/speech.
• model – Specifies the model to invoke, for example moss-tts; /v1 uses a single model field to name the model.
• delivery_method – audio returns the audio directly; url returns JSON with a result URL.
• response_format – The audio encoding format, such as mp3 / wav; defaults to mp3.
• async / webhook_url – Pass async=true to create an asynchronous task; once it reaches a terminal state, the result is delivered to webhook_url.
• task_id – The server returns a task_id for asynchronous tasks, used to query status and results.

Endpoint catalog