Voiceflow Mcp Server Client

Voiceflow Mcp Server Client is an MCP server that acts as a Node.js client connecting external Model Context Protocol servers to Voiceflow conversational AI agents. Built primarily for conversational designers, bot developers, and workflow automation engineers, it serves as a central proxy that aggregates multiple MCP capabilities into an accessible interface for Claude models. It communicates with local and remote MCP servers including Playwright, Brave Search, Google Calendar, weather services, and Zapier over HTTP or Server-Sent Events. The server orchestrates tool discovery, manages recursive function calling, formats responses, and handles execution errors automatically. By exposing these integrated endpoints to Voiceflow, it allows conversational chatbots to query search engines, check schedules, automate external browser interactions, and trigger third-party workflows during customer conversations without requiring custom API middleware for each integration.

Category: AI & LLM Tooling

Tags: chatbots, conversational-ai, voiceflow, workflow-automation

Visit Voiceflow Mcp Server Client

How to install and configure Voiceflow Mcp Server Client

1. Clone the repository with submodules: git clone --recursive https://github.com/voiceflow-gallagan/voiceflow-mcp-server-client.git 2. Navigate to the directory and install dependencies: npm install 3. Build the bundled MCP servers: npm run build 4. Copy the environment file template: cp .env.template .env 5. Populate .env with your ANTHROPIC_API_KEY, BRAVE_API_KEY, and Google OAuth variables if needed. 6. Configure external servers inside servers-config.json defining command arguments and environment variables. 7. Start the service using npm start or run containerized via docker-compose up -d.

What you can do with Voiceflow Mcp Server Client

• Querying real-time web information and news via the Brave Search integration directly within a Voiceflow chatbot dialog. * Managing calendar events by enabling Voiceflow agents to read, schedule, and update Google Calendar bookings dynamically. * Running automated headless browser flows via Playwright to fetch dynamic web page contents during live customer chats. * Triggering Zapier automation workflows by connecting to remote MCP endpoints through SSE configuration in the environment file. * Aggregating multi-tool sequences with automatic follow-up steps and response truncation to prevent token context overflow.

Key facts

• https://github.com/voiceflow-community/voiceflow-mcp-server-client
• AI & LLM Tooling, CRM, ERP & E-commerce
• chatbots, conversational-ai, voiceflow, workflow-automation

Part of MCP Servers

How do I install Voiceflow Mcp Server Client?

Clone the repository using git with recursive submodules, install required packages using npm install, and run npm run build to set up Playwright and submodules. Configure your credentials in a .env file based on .env.template, define server paths in servers-config.json, and launch the service via npm start or Docker.

What can Voiceflow Mcp Server Client do?

The client acts as a bridge connecting Voiceflow assistants to multiple Model Context Protocol servers. It handles tool discovery, executes tool calls using Anthropic Claude models, coordinates headless browsing via Playwright, integrates search APIs, connects with Google Calendar, and routes automation requests through Zapier.

Which tools are pre-configured in this client?

The client repository includes built-in configurations and build scripts for Playwright web automation, Brave Search, a time utility, an OpenWeather server, Google Calendar, and remote dynamic connections to Zapier over Server-Sent Events.

How do I configure servers in Voiceflow Mcp Server Client?

Define local or command-line MCP servers in the servers-config.json file by specifying the executable command, arguments, and required environment variables. For remote Zapier integration, set the ZAPIER_MCP_URL environment variable in your .env file to enable dynamic connection upon server startup.