RunningHub API icon

RunningHub API

Interact with the RunningHub API

Actions25

Standard Model API → Text To Audio Speech 2.8 Turbo

AI-generated

Summary

Submits a text-to-speech job to RunningHub's MiniMax Speech 2.8 Turbo endpoint (/openapi/v2/rhart-audio/text-to-audio/speech-2.8-turbo) using structured voice, emotion, speed, volume and pitch settings.

Inputs

  • Text (required) — Text to synthesize. Supports interjections such as (laughs), (sighs), (applause).
  • Voice ID — Built-in voice preset (e.g. Elegant_Man, Calm_Woman, Young_Knight). Default Elegant_Man.
  • Emotion — angry, disgusted, fearful, happy, neutral, sad or surprised. Default happy.
  • Speed — Playback speed multiplier, 0.5–2. Default 1.
  • Volume — Volume multiplier, 0.1–10. Default 1.
  • Pitch — Pitch offset in semitones, -12 to 12. Default 0.
  • Send Pronunciation Dictionary — When enabled, shows the Pronunciation Dictionary list.
  • Pronunciation Dictionary — List of override entries such as ASAP/As soon as possible (RunningHub allows up to 20). Entries are trimmed and blanks dropped. Sent as pronunciation_dict only if at least one entry remains.
  • Enable Base64 Output — Requests base64 audio in the response payload when supported (enable_base64_output).
  • English Normalization — Enables the model's English normalization option (english_normalization).

Output shape

One JSON item per input item, containing the RunningHub API response for the submitted job. This is typically a task record with a status and task ID.

The call is asynchronous. If the status is QUEUED or RUNNING, poll with Query Task Output V2 (/openapi/v2/query) to get the final audio URL. The node needs a RunningHub API key and a license key in the credentials. With Continue On Fail enabled, failed items return the input JSON plus an error.

Examples

Example 1: Narrate a generated article with a calm female voice

Text = {{$json.article}}, Voice ID = Calm Woman, Emotion = neutral, Speed = 1

Example 2: Read text with custom pronunciations

Enable Send Pronunciation Dictionary, add entry 'ASAP/As soon as possible', Voice ID = Casual Guy

Example 3: Get inline audio data for downstream processing

Text = message, Enable Base64 Output = true, then poll with Query Task Output V2 if status is not final

Links

Discussion