Read-aloud engine for Windows (.NET Framework 4.8, no UI framework): grabs the selected text from any app (UI Automation / copy / pointer / clipboard), turns screen text into speech-ready text (symbols, bullets, ALL-CAPS, English words inside Turkish text, user pronunciation dictionary), speaks it with SAPI voices or the offline Turkish neural voice (EMA Lightning on ONNX Runtime) with word-level positions, reads text off the screen with the built-in Windows OCR, translates with MyMemory/DeepL, and ducks other audio. The engine behind the free Sesli Okuma tray app.
Offline text-to-speech on the command line, powered by Kokoro-82M. Speaks text aloud or renders it to a WAV file, with 157 voices across nine languages and on-the-fly voice blending. Model weights and voices are downloaded on first use and cached; nothing leaves your machine.
This is a client library for the AivisCloud API. It allows you to easily integrate text-to-speech functionality into your .NET applications.
Text-to-speech (TTS) for .NET via the AI/ML API. Generate natural speech audio from text with multiple voices and formats. Supports streaming audio, saving to file, and returning byte arrays. Powered by AI/ML API.
HTTP client for OmniVoiceApi — multilingual zero-shot text-to-speech (voice cloning). Targets GET /health, POST /voices, POST /tts.
Idiomatic, async-first .NET client for the Fish Audio API: text-to-speech (buffered and streaming), realtime WebSocket TTS, speech-to-text, and voice-model management. Hand-crafted, MessagePack-correct, live-verified against the real API.
Small unofficial metadata and URL helpers for Seed Audio AI at seedaud.io.
Text read aloud by Apple's voices (Eloquence and the rest) and sped up by speechwarp, rendered a little ahead of playback, for iOS. Not an official Apple or Google product.
Connector to the ElevenLabs API: text to speech returning an audio file ready to be saved, the voices of the account and the subscription usage. The API key comes from the platform settings (ElevenLabsApiKey) or from a host-provided key provider.
CosyVoice importer for Sharplight: reads FunAudioLLM's Fun-CosyVoice3 text-to-speech release (Qwen2 speech-token language model, DiT flow, causal HiFT vocoder, the ONNX S3 tokenizer v3 and CAM++ speaker encoder) into Sharplight speech-synthesis models.
Sharplight with its defaults: references the core, the built-in architecture adapters, every model importer (GGUF, Hugging Face, ONNX, whisper.cpp, diffusers, the text-to-speech, voice activity and Ultralytics formats, .slm), the CPU, Metal and Vulkan execution providers and the native audio and video libraries, and wires them up with services.AddSharplight() (the task layer included).