sbbhimji/tts-mcp-server

Text-to-Speech MCP Server for SDLC Progress Announcements

A Model Context Protocol (MCP) server that provides real-time audio feedback during the Software Development Lifecycle workflow using AWS Polly.

Overview

This TTS MCP server integrates with the CLI Agent Orchestrator to announce progress at key milestones throughout the SDLC workflow, providing audio feedback without requiring constant screen monitoring.

Features

• Real-time Progress Announcements: Audio notifications at each SDLC stage
• AWS Polly Integration: High-quality neural text-to-speech
• Cross-platform Support: Works on macOS, Linux (Windows via WSL)
• Configurable: Enable/disable via environment variable
• Multiple Voices: Support for different AWS Polly voices

Prerequisites

• Python 3.8+
• uv package manager installed (Installation guide)
• AWS account with Polly access
• AWS credentials configured (aws configure)
• IAM permissions for polly:SynthesizeSpeech
• Audio output device

Installation

The TTS MCP server is automatically installed and managed by uv when you install the supervisor agent. No manual installation or virtual environment setup is required.

Update supervisor.md with Absolute Path

Edit supervisor.md and replace the path with your absolute path:

tts-mcp-server:
  type: stdio
  command: uv
  args:
    - "run"
    - "python"
    - "/Users/yourusername/echo-reinvent/echo-architect/tts-mcp-server/server.py"

Install Supervisor Agent

cao install ../cli_agent_orchestrator/agent_store/supervisor.md

Usage

Launch Supervisor with TTS

cao launch --agents supervisor

Give It a Task

"Create a social media app like Facebook"

You'll hear announcements like:

• 🔊 "Starting new project social-media-app"
• 🔊 "Creating specifications for your project"
• 🔊 "Specifications completed. Beginning parallel development phase"
• 🔊 "Building AWS solution with CDK and React"
• 🔊 "Generating architecture diagrams"
• 🔊 "Calculating AWS cost estimates"
• 🔊 "Creating user stories in Jira"
• 🔊 "All parallel tasks completed"
• 🔊 "Publishing artifacts to GitHub repository"
• 🔊 "Project completed successfully. Total time: X minutes"

Configuration

Enable/Disable TTS

# Enable (default)
export ENABLE_TTS=true

# Disable
export ENABLE_TTS=false

Available Voices

• Joanna (Female, US) - Default, general progress
• Matthew (Male, US) - Error notifications
• Salli (Female, US) - Completion messages

Full list of AWS Polly voices

Troubleshooting

No Audio Output

Check AWS credentials:

aws sts get-caller-identity

Verify Polly access:

aws polly describe-voices --language-code en-US

Test audio system:

# macOS
afplay /System/Library/Sounds/Ping.aiff

# Linux
speaker-test -t wav -c 2

TTS Not Working

Check environment:

echo $ENABLE_TTS  # Should be 'true' or empty

Reinstall supervisor:

cao install ../cli_agent_orchestrator/agent_store/supervisor.md

Permission Errors

Add Polly permissions to your IAM user/role:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": "polly:SynthesizeSpeech",
      "Resource": "*"
    }
  ]
}

Module Not Found Error

Verify the absolute path in supervisor.md is correct:

cat ../cli_agent_orchestrator/agent_store/supervisor.md | grep server.py

Reinstall supervisor agent:

cao install ../cli_agent_orchestrator/agent_store/supervisor.md

Files

tts-mcp-server/
├── server.py              # FastMCP TTS server
├── pyproject.toml         # Python project configuration for uv
├── requirements.txt       # Python dependencies
├── .gitignore            # Git exclusions
└── README.md             # This file

Important Notes

• ✅ No virtual environment needed - uv manages dependencies automatically
• ✅ Use absolute paths - Update supervisor.md with your full path to server.py
• ✅ Dependencies auto-installed - uv handles everything when supervisor launches