Narrate a listing in your avatar's voice: TTS with avatar_handle

To narrate a property listing in the same voice as your avatar, pass avatar_handle to POST /v1/tts-1.0/generate: Sume resolves the avatar's voice at submit time. A 1,200-character script costs $0.057 at $0.0475 per 1,000 characters.

Cost by script length

TTS charges by character, spaces and punctuation included. Spoken length depends on the voice and pace, so measure the audio you get back.

Request

Pick an avatar from GET /v1/avatar-1.0/avatars whose voice.status is ready. Ask for wav if you will mux or time the audio later; the default is mp3. Do not send model: TTS 1.0 has no engine picker and rejects it.

curl -X POST https://api.sume.com/v1/tts-1.0/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: listing-12-oak-street-voice" \
  -d '{
    "transcript": "Welcome to this bright three-bedroom home with a south-facing garden.",
    "avatar_handle": "agent_host",
    "output_format": { "container": "wav", "sample_rate": 44100 },
    "timestamps": { "words": true }
  }'

Limits and gotchas

• The transcript is up to 20,000 characters; audio over 1,200 seconds fails with tts_duration_exceeded and is not charged.
• If the avatar voice is not ready, the call fails, so list avatars first.
• State facts about a listing only if the agent has verified them; the voice reads the text as written.