Echogarden is a comprehensive, cross-platform speech processing toolset for Node.js developers and command-line users. It offers a unified interface to a wide array of speech technologies, including synthesis, recognition, translation, alignment, and noise reduction, aiming to simplify complex audio workflows without requiring Python or Docker.

How It Works

Echogarden leverages a modular architecture, integrating various speech engines through pure TypeScript, WebAssembly ports, or ONNX runtime. This approach avoids platform-specific binaries and allows for flexible use of both local, high-quality models (like VITS and Whisper) and cloud-based APIs from major providers. Its internal package system manages resource downloads, simplifying setup and model management.

Quick Start & Requirements

• Install globally: npm install -g echogarden@latest
• Requires Node.js v18 or later (v22+ recommended).
• Official documentation: https://echogarden.dev/

Highlighted Details

• Supports Text-to-Speech with Kokoro and VITS offline models, plus 16 other engines (including cloud providers).
• Features a TypeScript/ONNX port of Whisper for Speech-to-Text, supporting 100+ languages with word-level timing.
• Includes advanced features like speech-to-transcript alignment, speech translation, language detection, voice activity detection, and speech denoising.
• Offers subtitle generation with sentence boundary awareness and TTS pronunciation enhancements.

Maintenance & Community

• Developed by Rotem Dan.
• Community links and roadmap are available on the project website.

Licensing & Compatibility

• Licensed under GNU General Public License v3 (GPL v3).
• GPL v3 is a strong copyleft license, potentially restricting integration into closed-source commercial applications without specific licensing agreements or source code availability.

Limitations & Caveats

The GPL v3 license may impose significant restrictions on commercial use within proprietary software. While it supports ONNX Runtime, enabling CUDA execution providers for GPU acceleration requires specific configuration steps detailed in the documentation.