Multi-Source Documentation

You can run a dedicated MCP server that aggregates documentation from multiple sources so AI assistants can search, retrieve, and walk you through Voiceflow and Claude Code content. This server indexes docs, provides full‑text search, and exposes tools to find specific documents or code examples, making it easy to query and explore documentation from a single endpoint.

How to use

Connect your MCP client or Claude Desktop to the local MCP server that you run on your machine. Once connected, you can issue natural language queries to search across all indexed sources, retrieve specific documents by path or identifier, and drill into code examples or integration guides. Practical usage patterns include: asking for full‑text searches across Voiceflow or Claude Code docs, requesting step‑by‑step tutorials, or asking for documentation on a particular feature or API block. For example you can say you want to search for API blocks in Voiceflow docs, find code examples for integrating with external APIs, or show step‑by‑step guides for setting up a voice assistant.

How to install

Prerequisites: Python 3.10 or higher and the uv package manager.

Step by step installation and startup:

Clone the MCP server repository and navigate into the project folder.

Install dependencies using the package manager.

Run the MCP server directly with the runtime command.

Configuration

{
  "mcpServers": {
    "voiceflow_docs_mcp": {
      "command": "uv",
      "args": ["run", "voiceflow-docs-mcp"]
    }
  }
}

Notes on installation and running the server

You can start the server in one of two ways: run the MCP server directly with the runtime command or run it as a Python module. Use the method that fits your workflow.

Run directly from the command line to start the MCP server.

Alternative: start the server as a Python module.

Claude Desktop integration details

To connect Claude Desktop to this MCP server, add a configuration mapping that points Claude to the local runtime. The example below is provided as a reference configuration that you place in your Claude Desktop config file.

{
  "mcpServers": {
    "voiceflow_docs_mcp": {
      "command": "uv",
      "args": ["run", "voiceflow-docs-mcp"],
      "cwd": "/absolute/path/to/voiceflow-docs-mcp"
    }
  }
}

Available tools and how they help

The server exposes six specialized tools to work with documentation. Each tool enables a different way to search, locate, and retrieve information from the indexed sources.

Troubleshooting and tips

If you cannot connect, ensure the MCP server process is running, and confirm the client is configured to reach the local endpoint used by your setup. Check that indexing completed successfully on first run and that the correct sources (Voiceflow and Claude Code) are enabled. If you modify the data sources, restart the server to trigger reindexing.

Available tools

search_documentation

Performs a full-text search across all documentation sources with relevance ranking to surface the most relevant documents.

get_document

Retrieves a specific document by its exact path or identifier.

search_code_examples

Searches for code snippets and examples across the documentation sources.

list_categories

Lists available documentation categories and topics to help you navigate the content.

get_integration_docs

Fetches integration-specific documentation and guides.

find_step_documentation

Finds step-by-step tutorials and walkthroughs.