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

# Google Gemini Audio

> TTS and STT with Google Gemini

Audio processing using Google's Gemini models.

## Setup

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
export GOOGLE_API_KEY=your-key
```

## Text-to-Speech

```python theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
from praisonaiagents import AudioAgent

agent = AudioAgent(llm="gemini/gemini-2.5-flash-preview-tts")
agent.speech("Hello world!", output="hello.mp3")
```

## Speech-to-Text

```python theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
from praisonaiagents import AudioAgent

agent = AudioAgent(llm="gemini/gemini-2.0-flash")
text = agent.transcribe("audio.mp3")
print(text)
```

## Models

| Model                                 | Type |
| ------------------------------------- | ---- |
| `gemini/gemini-2.5-flash-preview-tts` | TTS  |
| `gemini/gemini-2.0-flash`             | STT  |