Characters that whisper, laugh and sound like they mean it.

Write cues like {sad}, {sighs} or {whisper} into your character’s lines and the v2 voices act them out. Give every character a voice of its own from a short recording.

Get an API key   › Emotion and delivery docs   ›

Hear it: characters cloned on v2, acting out their cues

  • {cues} for emotion, sounds and pauses
  • 40 languages, one voice
  • 3–30 s of audio to clone a voice

How it works

  1. 01

    Write the line

    Your LLM writes the reply with cues in curly braces. They’re never spoken.

  2. 02

    Choose the voice

    Use a v2 voice, or clone one from a 3 to 30 second recording.

  3. 03

    Stream it back

    Audio streams while it’s generated, so the character answers right away.

curl https://api.iodine.dev/v1/tts \
  -H "Authorization: Bearer $IODINE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"text": "{sad} You came back. {sighs} I missed you. {whisper}Stay a little longer.{/whisper}", "voice": "mia"}' \
  --output reply.wav

What you get

What it costs

One user talking for 20 minutes a day, with the character speaking for half of it, for a month.

300 minutes of speech on Iodine $2.25

The same at ElevenLabs v4 Turbo's published price $12.00

At Iodine’s pay-as-you-go prices. Other providers at their published pay-as-you-go rates; see the price comparison.

Questions

The v2 voices (Mia, Chloe, Zander and Ethan) and every voice you clone. On v1, cues are removed before speaking, except {laughs}.

Yes. v2 voices speak 40 languages: write the text in the language you want and send its code as language, such as hi for Hindi or ko for Korean.

Yes, with their permission. Clone from a 3 to 30 second recording and confirm the speaker consented; you must be able to show that permission if we ask.

Related

Start building with Iodine.

Top up from $5. No plan, seats or minimums.

Get an API key   ›
iodine