About MCP Expert Server
MCP Expert Server is an MCP server for teams working with API documentation. It uses Claude to generate queries from natural-language requests and answer questions using supplied documentation.
Key features
- 01Natural-language query generation
- 02Documentation question answering
- 03Loads TXT, Markdown, and JSON docs
- 04Customizable prompt files
Use cases
- 01Generate API queries from plain-language requests
- 02Find authentication guidance in API docs
MCP support
Exposes `create-query` and `documentation` tools to MCP clients. It uses Claude through an Anthropic API key to process requests and documentation.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEMCP Expert Server
A Model Context Protocol server that provides intelligent query generation and documentation assistance using Claude AI. The server analyzes your API documentation and provides two main tools:
- create-query: Generates queries based on natural language requests
- documentation: Provides relevant documentation information based on questions
Prerequisites
- Node.js >= 18
- An Anthropic API key for Claude
Installation
- Clone the repository
- Install dependencies:
npm install
- Create a
.envfile with your Anthropic API key:
ANTHROPIC_API_KEY=your_api_key_here
Setup
Before running the server, you need to:
- Build the project and run the setup script:
npm run build
npm run setup
This will:
- Create the required directories (
docs/andprompts/) - Create default prompt files
- Generate an initial service description
-
Add your API documentation files to the
docs/directory (supports.txt,.md, and.jsonfiles) -
Optionally customize the prompts in the
prompts/directory:system-prompt.txt: Main system prompt for Claudetool-metadata.txt: Additional context for tool descriptionsquery-metadata.txt: Additional context for query generationservice-description.txt: Auto-generated service description
Usage
Standalone Server
Start the server:
npm start
The server exposes two tools via the Model Context Protocol:
-
create-query: Generate a query based on natural language request
{ "name": "create-query", "arguments": { "request": "Find all users who signed up in the last week" } } -
documentation: Get information from the documentation
{ "name": "documentation", "arguments": { "request": "How do I authenticate API requests?" } }
Claude Desktop Integration
- Add this configuration to your Claude Desktop config file:
{
"mcpServers": {
"expert": {
"command": "node",
"args": ["/ABSOLUTE/PATH/TO/expert-server/build/index.js"],
"env": {
"ANTHROPIC_API_KEY": "your_api_key_here"
}
}
}
}
-
Replace
/ABSOLUTE/PATH/TO/expert-serverwith the actual absolute path to your server installation. -
Restart Claude Desktop.
Directory Structure
.
├── docs/ # Your API documentation files
├── prompts/ # System prompts and metadata
│ ├── system-prompt.txt # Main system prompt
│ ├── tool-metadata.txt # Tool description context
│ ├── query-metadata.txt # Query generation context
│ └── service-description.txt # Generated service description
├── src/ # Source code
│ ├── index.ts # Entry point
│ ├── server.ts # MCP server implementation
│ └── services/ # Core services
│ └── expertService.ts # Claude integration
└── package.json
Development
- Build the project:
npm run build
- The server uses TypeScript and follows a modular architecture
- All Claude interactions are handled by the ExpertService class
- Debug logs are written to stderr with [DEBUG] prefix
Troubleshooting
If you encounter connection issues:
- Ensure you've run the setup script:
npm run setup
- Check that all required files exist in the
prompts/directory - Verify your
ANTHROPIC_API_KEYis correctly set - Use absolute paths in your Claude Desktop config
- Check the debug logs (written to stderr)
Environment Variables
ANTHROPIC_API_KEY: Your Anthropic API key (required)
License
MIT
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.