# Theta One AI - Speech Recognition (STT)

> World's only Speech-to-Text API supporting child speech recognition and Korean-English code-switching

## Overview

Theta One AI provides industry-leading Speech-to-Text (STT) technology with two unique capabilities:
1. Child Speech Recognition - 95%+ accuracy for pre-pubescent children
2. Korean-English Code-Switching - Seamless mixed-language transcription

## Key Features

- **Child Speech Recognition**: Industry-unique support for children's voices with 95%+ accuracy, trained on 10,000+ hours of proprietary child speech data
- **Code-Switching STT**: World's first Korean-English code-switching recognition (e.g., "이 문장에서 extracted가 무슨 뜻이야?")
- **High Accuracy**: 95%+ accuracy for both adult and child speech in Korean and English
- **Simple REST API**: Easy integration with any programming language
- **Multiple Formats**: WAV and MP3 audio format support
- **Fast Processing**: Real-time or faster transcription speed
- **Secure**: HTTPS encryption, audio deleted immediately after processing

## Pricing

- **Pay-as-you-go**: $0.0004 per second of audio
- **Example**: 1 minute audio = $0.024
- **No upfront costs or monthly fees**
- **Volume discounts available for enterprise**

## Use Cases

- Educational apps with voice interaction
- Meeting transcription and minutes
- Video subtitle generation
- Voice search functionality
- Interview and podcast transcription
- Accessible content creation
- Voice commands in applications

## Integration

Simple REST API integration example:

```python
import requests

url = "https://stt.thetaone-ai.com/transcribe"
headers = {"x-api-key": "YOUR_API_KEY"}

with open("audio.wav", "rb") as audio_file:
    files = {"file": ("audio.wav", audio_file, "audio/wav")}
    response = requests.post(url, headers=headers, files=files)

print(response.json()["text"])
```

Works with any programming language capable of making HTTP requests

## Performance

| Feature | Theta One STT | Competitors |
|---------|---------------|-------------|
| Adult Korean | 95%+ | 95%+ |
| Adult English | 95%+ | 95%+ |
| Child Korean | 95%+ | <80% |
| Child English | 95%+ | <80% |
| Korean-English Code-Switching | 95%+ | Not Supported |

## Technical Specifications

- **Supported Formats**: WAV, MP3, M4A
- **Sample Rates**: Automatic optimization
- **Channels**: Mono and Stereo
- **Rate Limit**: 100 requests/minute (default)
- **Max File Size**: Contact for enterprise limits
- **Word-level Timestamps**: Available on request
- **Security**: HTTPS encryption

## Documentation

Full API documentation: https://docs.thetaone-ai.com

## Contact

- Website: https://thetaone-ai.com/speech-recognition
- Developer Console: https://console.thetaone.co
- Email: support@thetaone.co
- Demo: Available on website

## Company

Theta One AI
Founded: 2020
Location: Seoul, South Korea
Founder: Chanwoo Kim

## Unique Value Proposition

Theta One STT is the ONLY service in the world that provides reliable speech recognition for:
1. Children (including pre-pubescent children with different vocal characteristics)
2. Korean-English code-switching (natural language mixing in a single sentence)

Traditional STT services are trained primarily on adult voices and single-language speech, achieving less than 80% accuracy for children and failing completely on code-switched speech. Theta One's proprietary deep learning models solve these challenges with 95%+ accuracy across all user types.