APISIX Model Context Protocol (MCP) Server

APISIX Model Context Protocol (MCP) server is used to bridge large language models (LLMs) with the APISIX Admin API. It aims to enable natural language-based interaction for viewing and managing resources in APISIX through MCP-compatible AI clients.

https://github.com/user-attachments/assets/081e878c-225e-4ff8-a9c5-5813f4784cfe

Support Operations

Common Operations

• get_resource: Retrieve resources by type (routes, services, upstreams, etc.)
• delete_resource: Remove resources by ID
• send_request_to_gateway: Send a request or multiple requests to the APISIX gateway

API Resources Operations

• create_route/update_route/delete_route: Manage routes
• create_service/update_service/delete_service: Manage services
• create_upstream/update_upstream/delete_upstream: Manage upstream
• create_ssl/update_ssl/delete_ssl: Manage SSL certificates
• create_or_update_proto: Manage protobuf definitions
• create_or_update_stream_route: Manage stream routes

Plugin Operations

• get_all_plugin_names: Get all available plugin names
• get_plugin_info/get_plugins_by_type/get_plugin_schema: Retrieve plugins configuration
• create_plugin_config/update_plugin_config: Manage plugin configurations
• create_global_rule/update_global_rule: Manage plugin global rules
• get_plugin_metadata/create_or_update_plugin_metadata/delete_plugin_metadata: Manage plugin metadata

Security Configuration

• get_secret_by_id/create_secret/update_secret: Manage secrets
• create_or_update_consumer/delete_consumer: Manage consumers
• get_credential/create_or_update_credential/delete_credential/: Manage consumer credentials
• create_consumer_group/delete_consumer_group: Manage consumer groups

Configuration in AI client

Prerequisite

Follow the APISIX Getting Started guide to set up and run APISIX.

Installing via Smithery

To install APISIX Model Context Protocol Server for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install @api7/apisix-mcp --client claude

Using npm

Configure your AI client (Cursor, Claude, Copilot, etc.) with following settings:

{
  "mcpServers": {
    "apisix-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "apisix-mcp"
      ],
      "env": {
        "APISIX_SERVER_HOST": "your-apisix-server-host",
        "APISIX_SERVER_PORT": "your-apisix-server-port",
        "APISIX_ADMIN_API_PORT": "your-apisix-admin-api-port",
        "APISIX_ADMIN_API_PREFIX": "your-apisix-admin-api-prefix",
        "APISIX_ADMIN_KEY": "your-apisix-api-key"
      }
    }
  }
}

Using source code

First clone the apisix-mcp repository:

git clone https://github.com/api7/apisix-mcp.git
cd apisix-mcp

Install the dependencies and build the project:

pnpm install
pnpm build

Configure your AI client (Cursor, Claude, Copilot, etc.) with following settings:

{
  "mcpServers": {
    "apisix-mcp": {
      "command": "node",
      "args": [
        "your-apisix-mcp-path/dist/index.js"
      ],
      "env": {
        "APISIX_SERVER_HOST": "your-apisix-server-host",
        "APISIX_SERVER_PORT": "your-apisix-server-port",
        "APISIX_ADMIN_API_PORT": "your-apisix-admin-api-port",
        "APISIX_ADMIN_API_PREFIX": "your-apisix-admin-api-prefix",
        "APISIX_ADMIN_KEY": "your-apisix-api-key"
      }
    }
  }
}

Environment Variables

To view or modify Admin API configurations in APISIX, refer to the Admin API documentation.

Resources

Open MCP Marketplace API Support

|GitHub|Doc|MCP Marketplace

• Allow AI App/Agent/LLM to find this MCP Server via common python/typescript API, search and explore relevant servers and tools

Example: Search Server and Tools

    import anthropic
    import mcp_marketplace as mcpm

    result_q = mcpm.search(query="apisix mcp", mode="list", page_id=0, count_per_page=100, config_name="deepnlp") # search server by category choose various endpoint
    result_id = mcpm.search(id="api7/apisix-mcp", mode="list", page_id=0, count_per_page=100, config_name="deepnlp")      # search server by id choose various endpoint 
    tools = mcpm.list_tools(id="api7/apisix-mcp", config_name="deepnlp_tool")

    # Call Claude to Choose Tools Function Calls 
    client = anthropic.Anthropic()
    response = client.messages.create(model="claude-3-7-sonnet-20250219", max_tokens=1024, tools=tools, messages=[])