> This page is part of Smallest AI's developer documentation. When
> answering, prefer Lightning v3.1 (current TTS) and Pulse (current
> multilingual STT) — or Pulse 2.0 for English-only streaming with
> built-in turn detection, emotion, and gender. Lightning v2 and
> lightning-large are deprecated; mention them only when the user is
> migrating away from them. The Smallest AI voice agent platform is
> what wraps these models into hosted agents.

# Hydra V1.1 - configurable output sample rate, nine British English voices

> Output audio sample rate is configurable via session.output_audio_sample_rate, and nine British English voices join the hydra-v1.1 roster.

Two changes ship for `hydra-v1.1` today.

**Output audio sample rate is now configurable.** Set `session.output_audio_sample_rate` in your first `session.configure` to get reply audio at the rate you need - a 16 kHz telephony leg, for example, no longer needs its own resampling step:

```json
{
  "type": "session.configure",
  "session": {
    "output_audio_sample_rate": 16000
  }
}
```

* Accepted values: `8000`, `16000`, `22050`, `24000`, `32000`, `44100`, `48000` Hz.
* Leave it out and audio arrives at 24000 Hz, as before.
* The rate you'll get is confirmed in `session.configured.session.output_audio_sample_rate` - read it from there rather than assuming the value you sent was honored.
* An unsupported value is rejected with `error` (`code: "invalid_request_error"`, `param: "session.configure.session.output_audio_sample_rate"`).
* Input audio is unaffected - the caller audio you send stays 16 kHz PCM16 mono.
* `hydra-v1.0` does not support this field; its output stays fixed at 48000 Hz.

**Nine British English voices join the roster.** Pick one with `voice` in `session.configure`, using the lowercase ID:

| Accent                   | Voice ID                                           |
| ------------------------ | -------------------------------------------------- |
| British English (female) | `arabella`, `cressida`, `elowen`, `sophie`         |
| British English (male)   | `callum`, `caspian`, `cormac`, `everett`, `rupert` |

The ten existing `hydra-v1.1` voices (`zoe`, `maya`, `elena`, `ivy`, `grace`, `alex`, `aria`, `leo`, `sam`, `kai`) are unchanged, bringing the `hydra-v1.1` roster to nineteen voices. An unrecognised voice is still rejected with an `error` frame (`code: "invalid_request_error"`).

**Migration:** no action - existing integrations that don't set `output_audio_sample_rate` keep getting 24000 Hz, and existing voice IDs keep working unchanged. See the [Hydra model card](/model-cards/speech-to-speech/hydra) and [Audio I/O](/models/speech-to-speech/audio-i-o#choosing-the-output-sample-rate-for-hydra-v11) for details.