Sample Nova Sonic Mcp

What is Sample Nova Sonic Mcp?

Sample Nova Sonic Mcp is a WebSocket-based bidirectional audio streaming application that integrates with Amazon Nova Sonic model for real-time speech-to-speech conversations. It runs on a Node.js server with a modern web client and is built for developers and users who want natural conversational AI with tool integration via MCP.

How to use Sample Nova Sonic Mcp?

Install Node.js v18.0.0 or higher, configure AWS credentials with Bedrock access, clone the repository, run npm install, then npm run build and npm start. Open http://localhost:3000 in a browser, grant microphone permissions, and click the call button to begin. MCP servers are configured through a mcp_config.json file. CLI mode is available via npm run cli -- --input=/path/to/audio.wav.

Key features of Sample Nova Sonic Mcp

• Bidirectional real-time audio streaming with low latency
• Multi-language interface (English and Chinese)
• Multiple AI voice personas (Tiffany, Matthew, Amy)
• MCP tool integration for extended functionality
• Audio visualization and barge-in interruption support
• CLI mode for processing pre-recorded audio files

Use cases of Sample Nova Sonic Mcp

• Real-time speech-to-speech AI assistant with tool access
• Multilingual customer service application
• Interactive voice applications with custom workflows
• Prototyping conversational AI using Amazon Nova Sonic and MCP

FAQ from Sample Nova Sonic Mcp

What does Sample Nova Sonic Mcp do differently from other speech AI tools?

It combines Amazon Nova Sonic’s speech-to-speech model with an MCP tool integration system, enabling real-time bidirectional audio streaming and the ability to call external tools during a conversation, all through a responsive web interface.

Which platforms and models does Sample Nova Sonic Mcp support?

It runs on any modern browser with WebAudio API support and requires a Node.js server (v18+). The speech model used is Amazon Nova Sonic via AWS Bedrock.

Does Sample Nova Sonic Mcp support MCP servers?

Yes. MCP servers can be configured in a mcp_config.json file, with support for command, arguments, environment variables, and enabling/disabling each server. The UI lists available tools per server.

What are the pricing and licensing terms?

The project is licensed under the ISC License. Usage requires an AWS account with Bedrock access; AWS charges apply for the Nova Sonic model and related services.

Are there any known limits or caveats?

Microphone permissions must be granted in the browser. A default 5-minute session timeout exists (adjustable in server.ts). WebSocket connections depend on correct AWS credential configuration.