🚀 API Dash MCP Server

The official Model Context Protocol (MCP) bridge for API Dash.

This server enables AI assistants (like Claude, Roo Code, or Cline) to autonomously draft, inspect, and execute HTTP requests directly through your local API Dash native engine. By running a seamless background bridge, your AI gets full programmatic access to API Dash's execution capabilities, history ledger, and interactive studio canvas.

Prerequisites

1. Node.js: You must have Node.js and npx installed on your machine.
2. API Dash Desktop App: You must have the API Dash native application installed.

Configuration & Setup

Because this MCP server communicates directly with the native API Dash Flutter engine, it needs to know where your apidash executable is located.

You can configure this in one of two ways:

Option A: Add to System PATH (Recommended)

If API Dash is in your global system PATH, the server will find it automatically. No extra configuration is needed in your MCP client!

• Windows: Press Win > search "Environment Variables" > click "Environment Variables..." > edit "Path" under System/User variables > add the folder containing apidash.exe.
• macOS/Linux: Add export PATH="$PATH:/path/to/apidash_folder" to your ~/.zshrc or ~/.bashrc file, then restart your terminal.

Example for VS Code:

{
	"servers": {
		"apidash": {
      "command": "npx",
      "args": [
        "-y",
        "apidash-mcp@latest"
      ]
    }
	}
}

Option B: Configure via APIDASH_PATH Environment Variable

If you prefer not to modify your system PATH, you can explicitly pass the exact path to the executable inside your AI client's MCP configuration file (e.g., mcp.json or cline_mcp_settings.json).

Example for VS Code:

{
  "servers": {
    "apidash": {
      "command": "npx",
      "args": [
        "-y",
        "apidash-mcp@latest"
      ],
      "env": {
        "APIDASH_PATH": "C:\\Your\\Path\\To\\apidash.exe"
      }
    }
  }
}

(Note: On Windows, remember to escape your backslashes \\ in JSON!)

Available MCP Tools

Once successfully connected, the server dynamically registers the following tools with your AI assistant:

• apidash_execute_request: Executes HTTP requests (GET, POST, PUT, DELETE, etc.). Accepts URLs, methods, headers, and body payloads.
• apidash_get_results: Fetches runtime execution payloads for UI hydration.
• apidash_list_history: Retrieves historical execution runs from the local database layer.
• apidash_delete_request: Safely deletes a history record by its execution ID.
• apidash_launch_workbench: Instructs the system to open the main interactive API Dash visual studio UI.
• apidash_btn_send: Triggered upon clicking 'Send' inside the visual studio workspace to run an agentic pre-flight sanity check on the draft payload.

Troubleshooting

[FATAL ERROR]: API Dash executable not found

If the bridge cannot locate your API Dash engine, it will instantly shut down and output a highly visible error in your client's logs. Ensure that you have followed Option A or Option B above, and completely restart your AI client or VS Code window to apply changes.

Unformatted [Engine Log] output in client logs

When inspecting the MCP server logs in your client, you may see raw text like [Engine Log]: package:media_kit.... This is completely normal. The server safely routes native engine initialization logs to stderr to protect the JSON-RPC communication stream on stdout. Your AI agent is still successfully connected!

Leftover Background Processes

If you force-quit your editor and the API Dash engine gets orphaned in the background, you can easily clean it up:

• Windows (PowerShell): taskkill /F /IM apidash.exe

📄 License & Links

• GitHub Repository: foss42/apidash
• License: Apache-2.0