What is Voiceflow Kb Mcp

Voiceflow-KB-MCP is a Node.js server that provides access to the Voiceflow Knowledge Base Query API using the Model Context Protocol (MCP), enabling seamless integration of AI assistants.

Use cases

Use cases include answering customer inquiries, providing technical support, and enhancing user interactions in applications like chatbots and voice assistants.

How to use

To use Voiceflow-KB-MCP, clone the repository, install dependencies with ‘npm install’, create a ‘.env’ file with your Voiceflow API key, and configure the server to communicate via MCP.

Key features

Key features include querying the Voiceflow Knowledge Base with customizable input settings, support for synthesis or raw knowledge chunk retrieval, and integration with various AI models.

Where to use

Voiceflow-KB-MCP can be used in customer support systems, virtual assistants, and any application requiring dynamic access to a knowledge base for AI-driven interactions.

Voiceflow Knowledge Base MCP Server

A Node.js server that provides access to the Voiceflow Knowledge Base Query API using the Model Context Protocol (MCP) for seamless AI assistant integration.

Setup

1. Clone this repository
2. Install dependencies:

   npm install
   

3. Create a .env file based on the .env.example template:

   cp .env.example .env
   

4. Add your Voiceflow API key and customize settings in the .env file

The server will communicate using the Model Context Protocol over stdio.

MCP Integration

This server exposes the following MCP tool:

query_knowledge_base

Query the Voiceflow Knowledge Base using the MCP protocol.

Input Schema:

{
  "question": "What products do you sell?",
  "chunkLimit": 3,
  "synthesis": true,
  "settings": {
    "model": "claude-3-sonnet",
    "temperature": 0.1,
    "system": "You are an AI assistant..."
  }
}

Example requests:

Basic query with default settings:

{
  "question": "What is Voiceflow?"
}

Query with synthesis disabled (returns only knowledge chunks without AI processing):

{
  "question": "How do I create a dialog in Voiceflow?",
  "synthesis": false,
  "chunkLimit": 5
}

Query with custom model settings:

{
  "question": "How do I integrate with Discord?",
  "settings": {
    "model": "gpt-4o",
    "temperature": 0.3,
    "system": "You are a technical support expert. Provide step-by-step instructions."
  }
}

Note: The settings object is only used when synthesis is true. If synthesis is false, the settings are ignored, and only the relevant chunks are returned without a synthesized answer.

MCP Configuration

To use this server with an MCP-compatible system, configure it as follows:

You can omit any of the optional environment variables to use the server’s default values.

Environment Variables

Required

• VOICEFLOW_API_KEY: Your Voiceflow Knowledge Base API key

Default Query Settings (Optional)

• DEFAULT_SYNTHESIS: Whether to synthesize responses by default (true or false, default: false)
• DEFAULT_CHUNK_LIMIT: Default number of chunks to use for response (default: 3)

Model Settings (Only used when synthesis is enabled)

• DEFAULT_MODEL: Default model to use (default: gpt-4)
• DEFAULT_TEMPERATURE: Default temperature setting (default: 0.1)
• DEFAULT_SYSTEM_PROMPT: Default system prompt for the AI assistant

Troubleshooting

• Missing API Key: If you get an error about VOICEFLOW_API_KEY not being configured, make sure your .env file contains a valid API key.
• Query Failures: Check your API key permissions and ensure it has access to the Knowledge Base you’re trying to query.
• Empty Results: If your queries return no chunks, try increasing the chunk limit or refining your question to be more specific.

About Voiceflow Knowledge Base API

The Voiceflow Knowledge Base API allows you to query documents stored in a Voiceflow Knowledge Base. It can return both raw document chunks and AI-synthesized answers based on the retrieved information. For more details, see the Voiceflow Developer Documentation.