Runpod Public Endpoints icon

Runpod Public Endpoints

Call Runpod Public Endpoints (sync/async)

Actions5

Generate Audio

AI-generated

Summary

Generate audio outputs using a selected Runpod Public Endpoint audio model by sending a JSON input specifying the request parameters.

Inputs

  • Model (required) — Choose the Runpod Public Endpoint model specialized for audio generation.
  • Input JSON (required) — A JSON object defining the input parameters for the selected audio model, passed under { input: ... }. The format depends on the model; for example, it may include audio URLs, language, and response format.
  • Download Output as Binary — If set to true, attempts to retrieve the output audio URL returned by the model and attach it as binary data output (currently returns JSON with the download URL).

Output shape

A single JSON object containing the model's response output merged with the model ID. If 'Download Output as Binary' is true and the output contains an audio URL, the JSON includes a 'downloadUrl' property with that URL.

The operation supports both synchronous and asynchronous execution modes, optionally waiting for completion with polling. Errors during generation throw node execution errors with relevant details. The input JSON must be well-formed and appropriate for the chosen model; defaults are provided if left empty.

Examples

Example 1: Generate transcription using Whisper audio model

Select 'Whisper V3 Large (Audio)' model; input JSON specifying the audio URL, language 'en', and response format 'json'; optionally enable download to get downloadUrl in output.

Links

Discussion