SayIt 1.0.6

SayIt

Zero-dependency .NET text-to-speech SDK powered by Microsoft Edge's free online voices. No API keys, no registration, no external dependencies.

中文文档 → README.zh.md

dotnet add package SayIt

🚀 Quick Start

using SayIt;

// One line — text to file
await SayIt.SaveAsync("你好世界", "hello.mp3");

// Text to audio stream
await using var audio = await SayIt.StreamAsync("Hello world");

// Specify voice
await SayIt.SaveAsync("Bonjour", "greeting.mp3",
    new SayItConfig().WithVoice("fr-FR-DeniseNeural"));

✨ Features

• Zero configuration — no API keys, no sign-up, no account
• Zero external dependencies — only uses System.Net.WebSockets and System.Text.Json
• 400+ neural voices across 100+ languages
• Cross-platform — .NET Standard 2.1
• Streaming — output to file, stream, or pipe
• Streaming input — feed text as IAsyncEnumerable<string> (e.g. LLM tokens)
• Adjustable — control rate, pitch, volume, output format

📖 API

⚡ Static API (one-liners)

🔧 Instance API (reusable config)

var speaker = new SayItSpeaker(
    new SayItConfig()
        .WithVoice("en-US-JennyNeural")
        .WithRate("+20%")
        .WithPitch("-2st"));

await speaker.SaveAsync("Hello", "hello.mp3");
await speaker.SaveAsync("World", "world.mp3");
await speaker.StreamToAsync("Hi", httpResponse.Body);
await foreach (var chunk in speaker.StreamChunksAsync("Hi"))
    // process chunk

⚙️ Configuration

var config = new SayItConfig()
    .WithVoice("zh-CN-XiaoxiaoNeural")     // Voice name
    .WithRate("+30%")                        // Speech rate
    .WithPitch("-4st")                        // Voice pitch
    .WithVolume("150%")                       // Volume
    .WithFormat(OutputFormat.Webm_24Khz_16Bit_Opus)
    .WithTimeout(60);                           // Timeout in seconds

🎵 Output Formats

• OutputFormat.Mp3_24Khz_96Kbps (default)
• OutputFormat.Mp3_24Khz_48Kbps
• OutputFormat.Mp3_48Khz_96Kbps
• OutputFormat.Webm_24Khz_16Bit_Opus
• OutputFormat.Pcm_16Khz_16Bit / Pcm_24Khz_16Bit / Pcm_48Khz_16Bit
• OutputFormat.Wav_16Khz_16Bit / Wav_24Khz_16Bit / Wav_48Khz_16Bit

📁 Samples

🔍 How It Works

SayIt uses the same WebSocket API that powers Microsoft Edge's "Read Aloud" feature. It connects to speech.platform.bing.com with an Edge-emulated User-Agent, sends SSML, and receives MP3 audio chunks in real time. No Azure subscription or API key required.

License

MIT