# Gemini 3.8 Flash TTS - Core.Today AI API

> Google Gemini 3.8 Flash TTS via Fal.AI. Expressive single-speaker narration with 30 prebuilt voices, delivery directions given as separate style instructions, and inline vocal events such as <laugh> and <sigh>. Returns WAV audio.

- **Provider**: Google
- **Model ID**: google/gemini-3.8-flash-tts
- **Category**: Audio & TTS
- **Credits**: 105 per 1000 characters of spoken text (0.104625 credits/char; style instructions not billed)
- **Speed**: Fast
- **Quality**: High

## Features
- 30 prebuilt Gemini voices
- Style instructions separate from the spoken text — and not billed
- Inline vocal events such as <laugh> and <sigh>
- Billed only on the characters actually spoken

## Use Cases
- Narration with a precisely directed tone (calm documentary, upbeat promo)
- Voice prompts and greetings for apps and assistants
- Podcast intros and short spoken segments
- Multilingual voice content generated from one pipeline

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Required
- **prompt**: string - Verbatim text to speak (billed per character). Put delivery directions in style_instructions; inline vocal events may use <laugh> or <sigh>. The complete input is limited to 8,192 tokens.

### Optional
- **style_instructions**: string - Delivery style, separate from the spoken transcript (e.g. "cheerful and friendly"). Not billed.
- **voice**: string (default: Kore) - Prebuilt voice for single-speaker speech.
  Options: Achernar, Achird, Algenib, Algieba, Alnilam, Aoede, Autonoe, Callirrhoe, Charon, Despina, Enceladus, Erinome, Fenrir, Gacrux, Iapetus, Kore, Laomedeia, Leda, Orus, Pulcherrima, Puck, Rasalgethi, Sadachbia, Sadaltager, Schedar, Sulafat, Umbriel, Vindemiatrix, Zephyr, Zubenelgenubi

## Examples

### Cheerful sign-off
A short line with a cheerful, friendly style instruction.
```json
{
  "model": "google/gemini-3.8-flash-tts",
  "input": {
    "prompt": "Have a wonderful day!",
    "style_instructions": "cheerful and friendly",
    "voice": "Kore"
  }
}
```

### Calm Korean documentary narration
Korean narration directed to sound slow and calm.
```json
{
  "model": "google/gemini-3.8-flash-tts",
  "input": {
    "prompt": "깊은 바다 아래, 빛이 닿지 않는 곳에서도 생명은 조용히 이어집니다.",
    "style_instructions": "calm, slow documentary narration with a warm low voice",
    "voice": "Charon"
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- Keep delivery directions in style_instructions — they are not billed, while everything in prompt is spoken and billed.
- Use inline events like <laugh> or <sigh> in prompt for natural reactions.
- Two-speaker dialogue is not exposed on this gateway yet; generate each speaker's lines as separate requests.
- Output is WAV; convert downstream if you need a compressed format.

## Documentation
https://fal.ai/models/google/gemini-3.8-flash-tts