Suno Voice

Create a custom voice persona from a short recording, for use as a singer on Suno music generation.

Overview

Suno Voice is a setup-time node — it does not execute as part of a workflow run. Instead, you configure it once via a 3-step modal that walks you through KIE.ai’s /voice/validate/voice/generate flow. Once setup completes, the node stores a voiceId and emits it at workflow runtime as the personaId input to Suno Generate / Suno Cover / Suno Extend, so any music those nodes produce sings in your custom voice.

The validation phrase is generated server-side per submission, so the verification recording cannot be prepared ahead of time — you must record yourself reading the exact phrase the server returns in Step 2.

Setup flow

The 3-step modal opens when you click Configure Voice on the node card.

  1. Source recording — upload an audio file (or paste a URL) and select the vocal segment to analyse with start/end seconds. Pick the phrase language. This kicks off /api/v1/voice/validate and polls /api/v1/voice/validate-info until KIE returns the verification phrase.
  2. Read & record — the modal displays a short phrase like “Harmonies fill the air with joyful melodies tonight”. Record yourself singing or speaking it, then upload the recording. Use the Regenerate phrase button if you want a different one.
  3. Voice details — fill in voice name, optional style (“Pop, female vocal”), description, and singer skill level. Click Create voice to call /api/v1/voice/generate and poll /api/v1/voice/record-info for the final voiceId.

The 20-credit charge is reserved when you click Create voice in Step 3, committed on success, and refunded on failure or timeout.

Configuration

Field Type Default Description
Source recording URL url Hosted audio clip or uploaded file containing your voice.
Vocal segment start integer (≥0) 0 Start time (seconds) of the vocal slice to analyse.
Vocal segment end integer (>start) 10 End time (seconds) of the vocal slice.
Language enum en Language for the validation phrase. One of en, zh, es, fr, pt, de, ja, ko, hi, ru.
Verify recording URL url Your reading of the validation phrase, uploaded to your library.
Voice name string (max 200) "" Display name for the voice.
Style string (max 500) "" Free-form description, e.g. “Pop, female vocal”.
Description string (max 500) "" Notes about the voice.
Singer skill level enum beginner beginner, intermediate, advanced, professional.

Inputs & outputs

Credits

Step Cost
/voice/validate, /voice/validate-info, /voice/regenerate 0 credits
/voice/generate (Step 3 “Create voice”) 20 credits

The 20-credit charge covers KIE’s validate + generate calls combined. KIE.ai does not publish per-call pricing for this flow, so the value is a conservative one-time default and may be tuned later via the model_pricing table.

Credits are reserved on POST /v1/suno/voice/generate and committed by the polling endpoint (GET /v1/suno/voice/record-info) when KIE reports status="success". On status="fail", credits are refunded to the original pools.

Best practices

Common use cases

Limitations