Kwaivgi Kling V2.6 Create Voice

Kwaivgi Kling V2.6 Create Voice

Playground

Kling 2.6 Create Voice is a model can generate custom voice. Upload an audio file to create a custom voice that can be used with the voice control feature in V2.6 video generation. The audio should be clean, noise-free, with a single voice, and duration between 5-30 seconds. Built for stable production use with a ready-to-use REST API, no cold starts, and predictable pricing.

Features

Kling v2.6 Create Voice is a lightweight helper endpoint for creating a reusable voice profile from an audio sample. The output is typically a voice identifier you can plug into Kling v2.6 “voice control” workflows (for example, generating dialogue in a video using your custom voice).

Use this when you want consistent narration or character speech across multiple Kling v2.6 generations, without re-uploading the same reference audio every time.

Key capabilities

• Create a reusable voice profile from audio Upload or link to a voice sample and get back a voice reference you can re-use across runs.

• Designed for Kling v2.6 voice control workflows The resulting voice can be used to drive speech generation in Kling v2.6 video endpoints that support custom voice IDs.

• Simple, single-input interface Minimal setup: provide a clean reference clip and you’re ready to create a voice.

• Supports common audio upload patterns Typically works with either a public URL or an uploaded audio file, depending on your integration.

• Better consistency across scenes Re-using the same created voice helps keep a stable vocal identity across multiple generations.

Parameters and how to use

• voice_url: (required) A URL (or uploaded file reference) pointing to the audio sample used to create the voice.

Media (Audio)

Provide a single voice sample that’s easy to learn from:

• Use a clean, single-speaker clip (no background music, no overlapping voices).
• Aim for consistent volume and minimal reverb/echo.
• If you want a specific style (e.g., calm narrator, energetic host), choose a sample that clearly matches that delivery.

After you finish configuring the parameters, click Run, preview the result, and iterate if needed.

Pricing

• $0.035 per run

Notes

How to write prompts that use a Voice ID

When you use Kling v2.6 video endpoints that support voice-controlled generation, you can reference created voices directly inside the text prompt.

• Prompt length limit: your positive prompt cannot exceed 2500 characters.
• Voice tag syntax: use <<<voice_1>>> (or <<<voice_2>>>) to specify which voice should speak.
• Voice order must match voice_list: <<<voice_1>>> refers to the first voice in the voice_list parameter; <<<voice_2>>> refers to the second voice.
• Up to 2 tones per task: a video generation task can reference at most 2 tones.
• Tone requires sound=on: when specifying a tone, the sound parameter must be on.
• Keep grammar simple: simpler sentence structure improves reliability. Example: The man <<<voice_1>>> said, “Hello.”
• Billing behavior: if voice_list is not empty and the prompt references a voice tag (e.g., <<<voice_1>>>), the task is billed using the “with voice generation” metric.
• Capability varies by mode/version: voice support differs across Kling model versions and video modes; check the current Capability Map for the endpoint you’re using.

Safety and permission

• Consent matters: only create voices from audio you own or have explicit permission to use.
• If the created voice sounds “off,” the fastest fix is usually a cleaner reference clip (single speaker, less noise, fewer artifacts).
• Keep voice creation and voice usage consistent: once you have a voice ID, re-use it rather than re-creating new voices for the same speaker.

Related Models

• Kling v2.6 Pro (Text-to-Video) – Use created voice IDs to generate videos with dialogue, ambience, and SFX.
• Kling v2.6 Pro (Image-to-Video) – Animate a still image into a video, optionally with voice-controlled speech.
• Kling Text-to-Audio – Generate sound effects and audio from text prompts.
• Kling Video-to-Audio – Generate or extract matching audio for an input video.

Authentication

For authentication details, please refer to the Authentication Guide.

API Endpoints

Submit Task & Query Result

set -euo pipefail

export WAVESPEED_API_KEY="your-api-key"

REQUEST_BODY=$(cat <<'JSON'
{
  "audio": "https://interactive-examples.mdn.mozilla.net/media/cc0-audio/t-rex-roar.mp3"
}
JSON
)

# 1. Submit the prediction.
SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  -X POST "https://api.wavespeed.ai/api/v3/kwaivgi/kling-v2.6/create-voice" \
  -H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "${REQUEST_BODY}")

TASK=$(printf '%s' "${SUBMIT_RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
PREDICTION_ID=$(printf '%s' "${TASK}" | jq -r '.id // empty')
if [ -z "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id
' >&2
  exit 1
fi
RESULT_URL=$(printf '%s' "${TASK}" | jq -r '.urls.get // empty')
if [ -z "${RESULT_URL}" ]; then RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"; fi

# 2. Poll until the prediction finishes.
while true; do
  RESPONSE=$(curl --silent --show-error --fail-with-body \
    "${RESULT_URL}" \
    -H "Authorization: Bearer ${WAVESPEED_API_KEY}")
  RESULT=$(printf '%s' "${RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
  STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')

  case "${STATUS}" in
    completed) printf '%s\n' "${RESULT}" | jq '.outputs'; break ;;
    failed|cancelled|timeout) printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
    created|processing) sleep 2 ;;
    *) printf 'Unexpected status: %s
' "${STATUS}" >&2; exit 1 ;;
  esac
done

Parameters

Task Submission Parameters

Request Parameters

Response Parameters

Result Request Parameters

Result Response Parameters