MCP Servers

模型上下文协议服务器、框架、SDK 和模板的综合目录。

IBM APIC MCP server exposes API Connect capabilities to your MCP clients and AI Agent workflows.

创建于 12/17/2025
更新于 about 12 hours ago
Repository documentation and setup instructions

apic-mcp-server

IBM APIC MCP server exposes API Connect capabilities to your MCP clients and AI Agent workflows.

mcp Using apic-mcp-server with MCP Clients

This MCP server can be integrated with various MCP clients such as Claude Desktop, VS Code, and Langflow etc.

💻 Suggested MCP Clients

Visual Studio Code Claude Desktop Langflow

✅ Prerequisites

  • Existing API Connect v10 or v12 customer either on SaaS or on-prem
    • Your provider organization name
    • API key to connect with you API Connect instance
    • Client ID configured on your API Connect instance
    • Client Secret configured on your API Connect instance
    • APIC Platform Endpoint of your instance
    • APIC Management Endpoint of your instance
  • APIC MCP server's npm package tar file OR the .mcpb installer
  • One of the suggested MCP Client application

📄 Debug Logs

When running the MCP server, INFO level logs are written to a daily rotating file in a directory called apic-mcp inside your home directory by default:

  • Windows:
    • Log files are located at %USERPROFILE%\apic-mcp\apic-mcp-YYYY-MM-DD.log
  • macOS / Linux:
    • Log files are located at ~/apic-mcp/apic-mcp-YYYY-MM-DD.log

The log directory is created automatically if it does not exist. Each log file contains one day's logs and up to 14 days of logs are retained. If the log directory cannot be created, logs are discarded.

Additionally, if you would like the log file to capture DEBUG level logs, update your chosen mcp client's env config to include the below env property:

LOG_LEVEL: 'debug'

For example:

    "servers": {
        "apic-mcp-server": {
            "command": "npx",
            "args": ["-y", "-p", "${input:tarPath}", "apic-mcp-server"],
            "env": {
                ...
                ...
                "LOG_LEVEL": "debug"
            }
        }
    }

📝 Setup Instructions for suggested MCP Clients

Visual Studio CodeVisual Studio Code and LangflowLangflow

  1. Choose the relevant template file for your client (mcp.vscode.json or mcp.langflow.json).
  2. Fill in your APIC configuration details in the template.
  3. Copy the configured mcp json
  4. Follow steps provided by the respective client to continue with the setup.

Claude Desktop Claude Desktop

Claude Desktop now supports easy installation through Anthropic's new Extensions feature. We provide a pre-packaged .mcpb file for seamless setup.

No manual configuration needed! Simply install the .mcpb and follow the setup wizard.

Step 1: Locate the Extension File

Find the .mcpb file you have obtained from this repository:

apic-mcp-server-x.x.x.mcpb

where x.x.x is the version of the APIC MCP server

Step 2: Install the Extension

  1. Simply double-click the .mcpb file
  2. Claude Desktop will automatically open and start the installation process

Step 3: Complete the Setup

  1. A setup wizard will appear asking for details (listed in the prerequisites section) related to the APIC instance you would like to connect to
  2. Click Save or Install
  3. Ensure you enable the extension, as claude sets it as disabled by default

Step 4: Start Using

Once installed, the API Connect tools exposed via the MCP server will be available in your Claude Desktop conversations.


🛠️ Available Tools

Currently available tools

|Service|Link| |---|---| |Analytics|Analytics tools doc|

More to come soon...

快速设置
此服务器的安装指南

安装命令 (包未发布)

git clone https://github.com/ibm-apiconnect/apic-mcp-server
手动安装: 请查看 README 获取详细的设置说明和所需的其他依赖项。

Cursor 配置 (mcp.json)

{ "mcpServers": { "ibm-apiconnect-apic-mcp-server": { "command": "git", "args": [ "clone", "https://github.com/ibm-apiconnect/apic-mcp-server" ] } } }