Kokoro MCP SERVER: Text To Speech (TTS)

A comprehensive Text-to-Speech toolkit built on Kokoro-82M with audio enhancement, Model Context Protocol (MCP) server integration, CLI interface, and Docker deployment.

📺 Demo Video

Features

• Kokoro-82M TTS Engine: Open-weight model with 82M parameters (510 tokens per pass)
• 🌐 Streamlit Web UI: Enterprise-grade management interface with real-time preview (OPTIONAL)
• Audio Enhancement: Professional processing with librosa (normalization, noise reduction, fade in/out)
• MCP Server: Model Context Protocol integration for Claude Desktop, Cursor, and other AI tools (OPTIONAL)
• CLI Interface: Command-line tools for quick generation (OPTIONAL)
• Batch Processing: Generate multiple audio files efficiently
• Script Processing: Convert complete video scripts with automatic text chunking
• Docker Support: Containerization with docker-compose
• Enterprise Features: Structured logging, configuration management, comprehensive testing
• CI/CD: GitHub Actions pipeline with automated testing

Streamlit Web Interface (Optional)

Beautiful web UI for managing all TTS functionality:

• 🎯 Single Generation - Convert text with real-time preview
• 📦 Batch Processing - Process multiple texts in one go
• 📄 Script Processing - Complete video script conversion
• 🔍 Voice Explorer - Compare all 12 voices side-by-side
• ⚙️ Configuration - Manage settings visually
• 📊 Analytics - Track generations with charts and statistics

Install with: pip install -e ".[streamlit]" or pip install -e ".[complete]"

Quick Start:

python run_streamlit.py
# Opens at http://localhost:8501

📚 See STREAMLIT_README.md for complete Streamlit documentation.

Building on Kokoro-82M

What Kokoro-82M Provides Out-of-the-Box: Kokoro-82M is an exceptional open-weight TTS model that delivers: core neural TTS inference with 82M parameters, a basic Python inference library (KPipeline), 10 professional voice packs (male/female, American/British), phonemization (G2P) system, and raw 24kHz audio output with a 510-token processing limit per pass.

What aparsoft-tts Adds: We integrate Kokoro-82M's excellent TTS inference with comprehensive development tooling and workflow enhancements. This toolkit adds:

1. Audio post-processing - Normalization, noise reduction, silence trimming, and fade in/out using librosa
2. Automated script workflows - Direct script-to-voiceover conversion with paragraph detection and gap management
3. IDE-native generation - MCP server integration eliminates context switching for Claude Desktop and Cursor users
4. Deployment infrastructure - Docker deployment, structured logging, configuration management, and comprehensive testing
5. Batch processing - CLI and Python APIs for processing multiple segments efficiently

Technical Implementation

Audio Enhancement (librosa Integration):

This toolkit adds an audio processing pipeline on Kokoro generated TTS output:

# Without enhancement - raw Kokoro output
audio = kokoro_pipeline(text)

# With enhancement
audio = enhance_audio(
    kokoro_output,
    normalize=True,        # Consistent volume
    trim_silence=True,     # Remove dead air
    noise_reduction=True,  # Spectral gating
    add_fade=True         # Smooth transitions
)

Result: Voiceovers ready for YouTube, podcasts, or content creation without additional audio editing.

**MCP Server In

... View full README on GitHub