<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> VoiceCraft is an AI-driven, free, open-source platform for bidirectional voice processing, integrating Microsoft Edge TTS (20+ Chinese voices) and SiliconFlow STT. It targets users and developers needing efficient voice conversion, offering a responsive web UI and OpenAI-compatible APIs deployed via Cloudflare Workers for global, zero-configuration access.

How It Works

Deployed on Cloudflare Workers for edge computing, VoiceCraft uses Microsoft Edge TTS for high-quality speech synthesis with customizable parameters (speed, pitch, style) and the SiliconFlow API for accurate speech-to-text transcription. A unified web interface allows seamless switching between TTS and STT modes, complemented by RESTful APIs mimicking OpenAI's format for programmatic integration.

Quick Start & Requirements

• Web Usage: Access https://tts.wangwangit.com.
• Local Deployment: Clone repo, install Wrangler CLI (npm install -g wrangler), run wrangler dev.
• Prerequisites: Node.js/npm for local deployment. Audio uploads limited to 10MB.
• Links: Demo: https://tts.wangwangit.com.

Highlighted Details

• Bidirectional Processing: Handles both Text-to-Speech and Speech-to-Text seamlessly.
• OpenAI TTS API Compatibility: Offers a RESTful API mimicking OpenAI's format.
• Serverless Edge Deployment: Cloudflare Workers ensure global distribution, high availability, and zero-configuration.
• Extensive Chinese Voices: Over 20 distinct Chinese TTS voices with adjustable parameters.

Maintenance & Community

• Contribution: Accepts Issues and Pull Requests.
• Community: WeChat public account "一只会飞的旺旺" for updates and support. No Discord/Slack mentioned.

Licensing & Compatibility

• License: MIT License.
• Compatibility: Permissive MIT license suitable for commercial use and integration into closed-source projects.

Limitations & Caveats

• STT Token: Programmatic STT may require a custom SiliconFlow API token; the web UI uses a default.
• Audio Upload Limit: STT audio files are capped at 10MB.
• External API Dependency: Relies on Microsoft Edge TTS and SiliconFlow, subject to their terms and availability.