Official ElevenLabs Model Context Protocol (MCP) server that enables interaction with powerful Text to Speech and audio processing APIs. This server allows MCP clients like Claude Desktop, Cursor, Windsurf, OpenAI Agents and others to generate speech, clone voices, transcribe audio, and more.

🚀 NEW: Enhanced Conversational AI Capabilities

⚡ CRITICAL: What We Added to the Original Server

This is an ENHANCED VERSION of the original ElevenLabs MCP server with COMPLETELY NEW conversational AI analysis capabilities:

🎤 BRAND NEW Conversational AI Tools:

• ✨ list_conversations - DIDN'T EXIST BEFORE: Get comprehensive information about ALL your conversational AI calls with complete metadata, filtering, and pagination
• ✨ get_conversation - DIDN'T EXIST BEFORE: Extract FULL TRANSCRIPTS with timestamps, speaker identification, and complete conversation analysis

📊 Before vs. After:

• ❌ Original Server: No way to access conversation data or transcripts from conversational AI
• ✅ Enhanced Version: Complete conversation analysis with full transcript extraction
• ❌ Original Server: Basic conversational AI creation without data access
• ✅ Enhanced Version: Comprehensive conversation management and analysis tools

🔧 Technical Enhancements:

• Official Client Integration: Replaced basic HTTP requests with official ElevenLabs Python SDK
• Comprehensive Error Handling: Added robust API integration and fallback mechanisms
• Full API Coverage: All conversational AI functionality now uses the official client library

🎯 Enhanced by YUV.AI

Special thanks to Yuval Avidani from YUV.AI for the comprehensive enhancement of conversational AI capabilities!

• Developer: Yuval Avidani
• Company: YUV.AI
• Links: Linktree | X/Twitter | Instagram

Yuval significantly enhanced this MCP server with advanced conversational AI tools that provide complete conversation analysis, comprehensive transcriptions, and detailed metadata extraction using the official ElevenLabs client library.

🚀 Features

Core Audio Tools

• Text-to-Speech: Convert text to high-quality speech with customizable voice settings
• Speech-to-Text: Transcribe audio files with speaker diarization support
• Voice Cloning: Create instant voice clones from audio samples
• Sound Effects: Generate custom sound effects from text descriptions
• Audio Isolation: Isolate speech from background noise
• Voice Design: Create and preview new synthetic voices

🎤 NEW: Conversational AI Analysis Tools ⚡ THESE ARE COMPLETELY NEW!

The following tools DID NOT EXIST in the original server and provide comprehensive access to all conversation data and transcriptions:

list_conversations ⚡ BRAND NEW TOOL

Get comprehensive information about all conversational AI calls - THIS FUNCTIONALITY NEVER EXISTED BEFORE

• Lists all conversations with complete metadata
• Filters by agent, time range, and success status
• Shows conversation IDs, agent details, duration, message counts
• Supports pagination with cursor-based navigation
• Parameters:
  • agent_id: Filter by specific agent
  • cursor: Pagination cursor
  • page_size: Results per page (1-100, default 30)
  • call_start_after_unix: Filter conversations after timestamp
  • call_start_before_unix: Filter conversations before timestamp
  • call_successful: Filter by success status (success/failure/unknown)

get_conversation ⚡ BRAND NEW TOOL

Get complete conversation details including full transcriptions - TRANSCRIPT ACCESS NEVER EXISTED BEFORE

• Retrieves comprehensive conversation information
• Full transcript extraction with timestamps and speaker identification
• Audio availability information (user audio, response audio)
• Conversation metadata (duration, start time, status)
• Automatic analysis and sentiment data when available
• Smart waiting for conversation completion
• Parameters:
  • conversation_id: The conversation ID to retrieve
  • wait_for_completion: Wait for ongoing conversations (default True)
  • max_wait_seconds: Maximum wait time in seconds (default 300)

Example transcript output:

=== FULL TRANSCRIPT (27 messages) ===
[00:00] AGENT: Hi, I'm Sarah. Before we get started, who am I speaking with?
[00:05] USER: John.
[00:08] AGENT: Nice to meet you, John. Could you briefly describe your role...

Agent Management

• Create Agents: Set up conversational AI agents with custom prompts and voices
• List Agents: View all available conversational AI agents
• Agent Details: Get comprehensive agent configuration and status

Voice Management

• Voice Library: Search and browse the complete ElevenLabs voice library
• Voice Search: Find voices by name, description, and categories
• Voice Details: Get detailed information about specific voices
• Voice Creation: Generate new voices from text descriptions

Advanced Features

• Phone Integration: Make outbound calls via Twilio
• Knowledge Base: Add documentation and context to agents
• Subscription Management: Monitor API usage and limits
• Audio Playback: Play generated audio files locally

🔧 Transport Protocol

This MCP server uses STDIO transport, providing real-time communication with MCP clients like Claude Desktop through standard input/output streams.

📋 Prerequisites

• Python 3.11+ (Required by ElevenLabs SDK)
• ElevenLabs API Key (Get one at elevenlabs.io)
• Windows (Batch scripts included for Windows)

🛠️ Installation

1. Clone the Repository

git clone https://github.com/hoodini/elevenlabs-mcp.git
cd elevenlabs-mcp

2. Set Up Virtual Environment

python -m venv .venv
.venv\Scripts\activate  # Windows
# source .venv/bin/activate  # Linux/Mac

3. Install Dependencies

pip install -e .

4. Configure Claude Desktop

Add to your Claude Desktop configuration file:

Location: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "ElevenLabs": {
      "command": "C:\\path\\to\\elevenlabs-mcp\\run_mcp_server.bat",
      "args": [],
      "cwd": "C:\\path\\to\\elevenlabs-mcp"
    }
  }
}

Replace C:\\path\\to\\elevenlabs-mcp with your actual project path.

5. Set Your API Key

Edit run_mcp_server.bat and replace the API key:

set ELEVENLABS_API_KEY=your_actual_api_key_here

🎯 Usage Examples

🎤 NEW: Conversational AI Analysis (THESE TOOLS ARE BRAND NEW!)

# List all recent conversations
Use the list_conversations tool to see all conversational AI calls

# Get detailed conversation with full transcript
Use get_conversation with a specific conversation_id to see:
- Complete conversation transcript with timestamps
- Audio availability information  
- Call metadata and duration
- Success/failure status

Text-to-Speech

# Convert text to speech
Use text_to_speech with your text and preferred voice

# Custom voice settings
Adjust stability, similarity_boost, and speed parameters

Voice Cloning

# Clone a voice from audio files
Use voice_clone with audio file paths and description

# Create voice from text description
Use text_to_voice to generate voice previews

🔍 MCP Inspector

Test your server setup with the MCP Inspector:

npx @modelcontextprotocol/inspector@latest

The inspector will help you:

• Verify server connection
• Test individual tools
• Debug configuration issues
• Explore tool capabilities

📊 API Coverage

This MCP server provides comprehensive coverage of the ElevenLabs API:

✅ Fully Implemented

• Conversational AI: Complete conversation management and transcript access ⚡ NEW ENHANCEMENT
• Text-to-Speech: All voice synthesis capabilities
• Speech-to-Text: Audio transcription with diarization
• Voice Management: Creation, cloning, and library access
• Agent Management: Full conversational AI agent lifecycle
• Sound Effects: Custom audio generation
• Audio Processing: Isolation and enhancement

🎯 Enhanced Features ⚡ ALL NEW ENHANCEMENTS

• Official Client Integration: Uses the official ElevenLabs Python SDK (NEW)
• Comprehensive Transcripts: Full conversation analysis with timestamps (COMPLETELY NEW)
• Advanced Filtering: Multiple filter options for conversation search (NEW)
• Error Handling: Robust error handling and fallback mechanisms (ENHANCED)
• Windows Integration: Optimized for Windows development environments (NEW)

🐛 Troubleshooting

Common Issues

"Module not found" errors:

• Ensure you're using the correct virtual environment
• Verify Python 3.11+ is installed
• Run pip install -e . in the project directory

MCP connection issues:

• Check that the batch file path is correct in Claude config
• Verify API key is set in the batch file
• Ensure no spaces in file paths cause parsing issues

Transcript not showing:

• Verify conversation is completed (status: done)
• Check that the conversation has audio content
• Some conversations may not have transcripts available immediately

📝 Development

Building the Package

python -m build

Running Tests

python -m pytest tests/

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Run tests and build
5. Submit a pull request

📄 License

MIT License - see LICENSE file for details.

🙏 Acknowledgments

• ElevenLabs for their excellent AI audio platform and API
• Anthropic for the Model Context Protocol specification
• YUV.AI & Yuval Avidani for the comprehensive conversational AI enhancements

🔗 Links

• ElevenLabs API Documentation
• Model Context Protocol
• YUV.AI Linktree
• Yuval Avidani X/Twitter

---

Made with ❤️ by the community, enhanced by YUV.AI