> ## Documentation Index
> Fetch the complete documentation index at: https://docs.zeroruntime.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Gradium

> Python API reference for the gradium speech-to-text provider.

## GradiumSTT

Realtime speech-to-text with Gradium, using your own Gradium key. Semantic voice
activity detection decides when the user has finished speaking.

Authenticates with GRADIUM\_API\_KEY from your environment, or an explicit
`api_key`.

```python theme={null}
from zeroruntime.plugins import GradiumSTT  # direct to the vendor, with your key
```

### Fields

**model** `str` (default: UNSET)

Realtime transcription model. Defaults to 'default'.

**language** `str` (default: UNSET)

Language of the audio: "en", "fr", "de", "es", "pt", or "any" to detect it. Defaults to 'en'.

**delay\_in\_frames** `int | None` (default: UNSET)

Audio context before text is emitted, in 80 ms frames. Lower is faster, higher is steadier. Leave unset for Gradium's default (10). Defaults to None (not set).

**temperature** `float | None` (default: UNSET)

Sampling temperature from 0 to 1.5. Leave unset for Gradium's default. Defaults to None (not set).

**keywords** `list[str] | None` (default: UNSET)

Up to 500 single-word names and terms to recognize, e.g. \["Gradium"]. Split multi-word names, and list every spelling you expect. Defaults to None (not set).

**vad\_horizon\_s** `float` (default: UNSET)

Which of Gradium's semantic VAD predictions ends a turn: how long the user must be expected to stay silent, 0.5, 1, 2 or 3 seconds. Longer is less likely to cut them off. Defaults to 2.0.

**vad\_threshold** `float` (default: UNSET)

Probability of that silence, from 0 to 1, that ends the turn. Defaults to 0.9.

**input\_sample\_rate** `int | None` (default: UNSET)

Override the detected input sample rate. Leave unset in normal use. Defaults to None (not set).

**input\_channels** `int | None` (default: UNSET)

Override the detected input channel count. Leave unset in normal use. Defaults to None (not set).

**base\_url** `str` (default: UNSET)

WebSocket endpoint of Gradium realtime speech-to-text. Change only for a proxy or a regional endpoint. Defaults to 'wss\://api.gradium.ai/api/speech/asr'.

**interim\_results** `bool` (default: UNSET)

Set to True to also send live partial transcripts to the client. Defaults to False.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.