OpenAI TTS Web Client

English | Русский | 中文

A simple unofficial web client for OpenAI Text-to-Speech API with multi-language interface support.

Features

• Automatic model loading: fetches all available TTS models via API
• 10 voices: alloy, ash, ballad, coral, echo, fable, nova, onyx, sage, shimmer
• GPT model support: with prompts for speech control (accent, emotions, tone)
• Flexible settings: speech speed, audio format, voice instructions
• Responsive design: works on desktop and mobile devices

Quick Start

1. Open https://cucumber148.github.io/openai-tts-client
2. Enter your OpenAI API key
3. Select model and voice
4. Enter text and click "Generate Speech"

Supported Models

The application automatically loads all available TTS models from your OpenAI account.

Audio Formats

• MP3 - Universal format (default)
• Opus - For internet streaming, low latency
• AAC - For mobile devices and YouTube
• FLAC - Lossless compression
• WAV - Uncompressed, low latency
• PCM - Raw audio data

Speech Control (GPT Models)

GPT models support instructions for controlling various aspects of speech:

• Accent: "Use a British accent"
• Emotions: "Speak cheerfully and energetically"
• Tone: "Speak like a radio host"
• Pace: "Speak slowly and clearly"
• Style: "Speak in a whisper"

Technical Features

• Automatic language detection based on browser settings
• Dynamic model support - automatically picks up new OpenAI models
• Local storage of API key and settings in browser
• Direct requests to OpenAI API without intermediate servers

Requirements

• OpenAI API key with access to TTS models

Limitations

• Maximum text length: 4096 characters
• Speech speed adjustable only for legacy models

Local Development

git clone https://github.com/cucumber148/openai-tts-client
cd openai-tts-client

Support

If you encounter issues:

1. Check browser console (F12) for errors
2. Verify API key is correct
3. Check OpenAI account limits
4. Create an issue in the GitHub repository

About the Project

I wrote this application in a few minutes for quick testing of the OpenAI TTS API. Decided to share it on GitHub in case it might be useful to someone.

License

GPL-3.0 License