home / mcp / webdna mcp server
Provides an MCP server that indexes WebDNA documentation and exposes search and retrieval tools via HTTP and stdin/stdout interfaces.
Configuration
View docs{
"mcpServers": {
"jacgood-webdna-mcp-server": {
"url": "http://localhost:3002/mcp",
"headers": {
"SUPABASE_KEY": "your-supabase-key",
"SUPABASE_URL": "https://your-supabase-url"
}
}
}
}You can run and query the WebDNA MCP Server to access, search, and retrieve WebDNA documentation through efficient MCP tools. It supports both an HTTP API for remote clients and a local stdin/stdout mode for integration with AI assistants, enabling fast, indexed access to WebDNA instructions and categories.
You interact with the server using an MCP client to invoke tools that search, fetch, and summarize WebDNA documentation. Start by ensuring the server is running in one of its supported modes, then call the available MCP tools to discover and retrieve documentation, or to run targeted queries by category, instruction ID, or random samples. Use the HTTP mode for remote integration and the stdin/stdout mode when integrating directly with AI assistants.
Prerequisites you need before installing are either a Docker-enabled environment or a Node.js development setup.
If you prefer containerized deployment, follow these steps to set up the MCP server with Docker.
If you opt to run directly without Docker, follow the sequence below to install dependencies, migrate data, scrape documentation into the database, and start the MCP server.
# Install dependencies
npm install
# Apply the database migrations to your Supabase project
# Use the Supabase SQL Editor to run the SQL in migrations/01_initial_setup.sql
```
```
# Populate the database by running the documentation scraper
npm run scrape
```
```
# Start the HTTP MCP server
npm run start:mcp-http
```
```
# Start the stdin/stdout MCP server for AI assistants
npm run start:mcpTo connect this MCP server to Continue.dev, provide the MCP endpoint and credentials in your project configuration so actions can be invoked directly from the editor. Include the following in your configuration to enable remote tool invocation and secure access to your Supabase-backed data store.
mcpServers:
- name: WebDNA Documentation
command: curl
args:
- -X
- POST
- http://localhost:3002/mcp/invoke_tool
- -H
- 'Content-Type: application/json'
- -d
- |
{"tool": "$0", "params": $1}
env:
SUPABASE_URL: "your-supabase-url"
SUPABASE_KEY: "your-supabase-key"Secure your MCP deployment by restricting access to the HTTP endpoint, using TLS in production, and managing secrets via environment variables or a secret manager. When using the Docker deployment, enable health checks and follow container security best practices to minimize resource usage in cloud environments.
If you encounter startup or connection issues, verify that the server is listening on the expected port (3002 for HTTP) and that the correct MCP command is used for the chosen mode. Check logs for errors related to database migrations, scraping, or tool invocations, and ensure your Supabase project is accessible with the provided URL and key.
Use the HTTP mode for integration with external clients and orchestration services. For local AI-assisted workflows, prefer the stdin/stdout mode to minimize latency between your assistant and the documentation data.
This server exposes a set of enhanced MCP tools to explore WebDNA documentation. You can search, fetch by ID, retrieve categories with counts, obtain random samples, and query overall statistics.
Searches WebDNA documentation with filtering and relevance scoring for queries across categories and instructions.
Retrieves detailed documentation for a specific WebDNA instruction or context by ID or name.
Lists all WebDNA documentation categories with counts of instructions in each category.
Returns a random subset of WebDNA documentation entries for exploratory browsing.
Provides statistics about the WebDNA documentation database, such as total entries and categories.