Apidog

The Apidog MCP Server provides an HTTP-based Model Context Protocol interface that lets AI tools discover and understand your API landscape. It exposes endpoints, schemas, and project metadata so you can search, inspect, and reuse API definitions inside your AI workflows.

How to use

You access the MCP server from an MCP-compatible client. Start by initializing the session, then use the available tools to explore your API surface. You can search for endpoints by keyword, retrieve detailed information about a specific endpoint, fetch model and schema definitions, and view project statistics for your APIs.

Practical usage patterns include: finding endpoints that match a topic you’re building against, inspecting a particular endpoint’s parameters and request/response shapes, loading the underlying schema for a model, and reviewing overall API project metrics to guide AI-assisted documentation or tooling.

How to install

Prerequisites you need before installing this MCP server are node-based runtime and a package manager.

Step-by-step commands you run locally:

pnpm install

Additional setup and configuration

Configure connection and access options to control who can use the MCP server and how it connects to your API sources.

PORT=3333

# Method A: Apidog project context (optional if you use a project)
APIDOG_ACCESS_TOKEN=your-token
APIDOG_PROJECT_ID=your-project-id

# Method B: Direct OpenAPI/Swagger URL
APIDOG_OAS_URL=https://petstore3.swagger.io/api/v3/openapi.json

# Optional: MCP server API key for authentication (if you want to enable access control)
MCP_API_KEY=your-mcp-api-key

Run and test

Start the server in development mode to test changes quickly. When you are ready for production, build the project and run the optimized build.

pnpm dev

Security and access control

If you enable authentication, provide your API key via the MCP_API_KEY value. Keep sensitive values out of source control and use environment variables or a secret manager in production.

Docker

You can run the MCP server inside a container to simplify deployment and isolation.

docker build -t apidog-mcp-server .
docker run -p 3333:3333 --env-file .env apidog-mcp-server

Available tools

apidog_search

Search API endpoints by keyword to quickly locate relevant endpoints for your task.

apidog_get_endpoint

Retrieve detailed information about a specific endpoint, including parameters and request/response schemas.

apidog_list_endpoints

List all endpoints available in the connected API sources for overview and discovery.

apidog_get_schema

Fetch model or schema definitions used by your APIs.

apidog_project_info

Provide project-level statistics and metadata about the API collection.

apidog_refresh

Refresh the internal cache to ensure you have the latest endpoint and schema data.