Quick Start

1. Initialize: Create a local MCP server instance.
2. Register: Load your CWL-described tools (supports local files or repositories).
3. Connect: Start the MCP server and link it to your MCP client.
4. Use Chat with the LLM agent to execute the task.

Basic Setup

The framework allows you to set up MCP servers with domain-specific tools described in CWL. The following example demonstrates how to create a server for PDF processing:

from coala.mcp_api import mcp_api
import os
base_dir = os.path.dirname(__file__)

mcp = mcp_api(host='0.0.0.0', port=8000)
mcp.add_tool(os.path.join(base_dir, 'pdftk_cat.cwl'))
mcp.add_tool(os.path.join(base_dir, 'pdf2docx.cwl'))
mcp.serve()

This creates an MCP server that exposes two command-line tools for PDF processing operations:

• pdftk_cat: The core command in the command-line tool PDFtk
• pdf2docx: The core command in the command-line tool pdf2docx

You can configure your MCP client (e.g., Cursor, Claude Code, Codex, Gemini CLI) to connect to the MCP server using one of two methods:

Option 1: Automatic Execution (Recommended)

In this mode, the MCP client launches the server automatically when it starts. This is generally the easiest method for local use.

{
    "mcpServers": {
        "PDF Processing": {
            "command": "python",
            "args": ["/path/to/examples/PDF/PDF-operation_mcp.py"]
        }
    }
}

Option 2: Connect to Running Server

Start the server by running the following command in your terminal:

python examples/PDF/PDF-operation_mcp.py

Once the server is running, you can connect your client through the following configuration:

{
    "mcpServers": {
        "PDF Processing": {
            "url": "http://localhost:8000/mcp",
            "transport": "streamable-http"
        }
    }
}

Start Using

With this setup, you can send natural language queries to the MCP Client (e.g., Cursor, Claude Code, Codex, Gemini CLI):

• "extract specified pages from xxx.pdf"
• "combine all the uploaded files into a single PDF document"
• "split the document at page xxx and save them as separate files"
• "convert all/specified pages of xxx.pdf into docx file"

The Client retrieves the tool list from the MCP server. The LLM selects the appropriate tool and sends a structured request for the analysis. Coala translates this selection into a CWL job and executes it within a container. The execution logs and results are returned to the LLM, which interprets them and presents the final answer to you.

Next Steps

• Check out use cases