Heygen Voice API Documentation

Heygen Voice API Documentation

Playground

HeyGen Voice Text-to-Speech converts text into expressive speech using public voices or reference recordings for instant voice cloning, supporting narration, voiceovers, multilingual content, digital humans, and voice production workflows. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

HeyGen Voice turns text into natural, expressive speech. Choose a public voice or provide a single-speaker recording for instant voice cloning, with language selection and adjustable expressiveness.

Why Choose This?

• 24 public voices — Choose a voice that fits your content.
• Instant voice cloning — Use a reference recording instead of a public voice.
• 39 supported languages — Select a language or let it be detected from the text.
• Expression control — Adjust the delivery with an expressiveness setting.
• WAV audio output — Generate speech as a 44.1 kHz WAV file.

Parameters

How to Use

1. Enter your text — Add the script you want spoken.
2. Choose a voice — Select a public voice or upload a single-speaker reference recording.
3. Adjust delivery — Optionally select the language and expressiveness.
4. Generate — Submit the request and download the speech audio.

Pricing

$0.04 per 1,000 input characters, calculated using the actual text length.

The same rate applies when using a public voice or a reference recording.

Best Use Cases

• Narration — Create spoken explanations and stories.
• Video voiceovers — Add speech to presentations and social content.
• Multilingual content — Generate speech in supported languages.
• Custom voices — Use a reference recording for voice consistency.

Pro Tips

• Use punctuation to guide phrasing and pauses.
• Provide a clear recording with one speaker for voice cloning.
• Set the language explicitly when the text is short or ambiguous.

Notes

• text is required and supports up to 5,000 characters.
• A reference recording takes priority over the selected public voice.
• Only the first 3 minutes of reference audio are used.
• Omit language to detect it from the text.
• Use only recordings you own or have permission to clone.
• The output is a WAV audio file, not a reusable voice ID.

Related Models

• HeyGen Avatar IV — Create talking-avatar videos from a portrait and speech audio.
• HeyGen Avatar V — Create talking-avatar videos from a supported avatar and speech audio.

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'
{
  "text": "A clear example input",
  "voice": "Jonah",
  "language": "ar",
  "expressiveness_boost": 1
}
JSON
)

# 1. Submit the prediction once. Never retry this request automatically:
# a retry after an unclear failure can create a second, billed task.
CHECK='The task may already exist. Check History or the predictions list before submitting again.'
if ! SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  --connect-timeout 10 --max-time 60 \
  -X POST "https://api.wavespeed.ai/api/v3/heygen/voice" \
  -H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "${REQUEST_BODY}"); then
  printf 'Submit failed. %s\n' "${CHECK}" >&2
  exit 1
fi

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 "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id. %s\n' "${CHECK}" >&2
  exit 1
fi
RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"

# 2. Poll until the prediction reaches a final status.
BODY_FILE=$(mktemp)
trap 'rm -f "${BODY_FILE}"' EXIT
while true; do
  # Transient errors (429, 5xx) are retried, honoring Retry-After.
  HTTP_CODE=$(curl --silent --show-error --fail-with-body \
    --connect-timeout 10 --max-time 60 --retry 5 --retry-delay 2 \
    --output "${BODY_FILE}" --write-out '%{http_code}' \
    "${RESULT_URL}" \
    -H "Authorization: Bearer ${WAVESPEED_API_KEY}" || true)
  case "${HTTP_CODE}" in
    2??) ;;
    000|429|5??)  # dropped connection or temporary error: keep polling
      printf 'Task %s: HTTP %s, polling again in 10s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      sleep 10; continue ;;
    *) printf 'Task %s: HTTP %s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      cat "${BODY_FILE}" >&2; exit 1 ;;
  esac
  RESULT=$(jq 'if type == "object" and has("data") then .data else . end' "${BODY_FILE}")
  STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')

  case "${STATUS}" in
    completed) printf '%s\n' "${RESULT}" | jq '.outputs'; break ;;
    failed|cancelled|timeout|deleted)
      printf 'Task %s ended with status %s\n' "${PREDICTION_ID}" "${STATUS}" >&2
      printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
    *) sleep 2 ;;
  esac
done

Parameters

Task Submission Parameters

Request Parameters

Response Parameters

Result Request Parameters

Result Response Parameters