RDL MCP Server

RDL MCP Server

mcp-name: io.github.bethmaloney/rdl-mcp

Edit SSRS reports using AI assistants instead of wrestling with 2000+ lines of XML. This Model Context Protocol (MCP) server gives Claude, Copilot, and other AI tools simple commands to read and modify RDL files.

What It Does

Read reports:

• describe_rdl_report - Get report structure overview
• get_rdl_datasets - View datasets, fields, and stored procedures (supports field limiting and filtering)
• get_rdl_parameters - List all report parameters
• get_rdl_columns - See column headers, widths, and bindings

Modify reports:

• update_column_header / update_column_width - Change columns
• add_column / remove_column - Add or remove columns
• update_column_format - Change number/date formatting
• update_stored_procedure - Swap stored procedures
• add_dataset_field / remove_dataset_field - Manage dataset fields
• add_parameter / update_parameter - Manage parameters
• validate_rdl - Validate XML after changes

Why it's better than editing XML:

• AI sees clean JSON instead of verbose XML namespaces
• One-line commands instead of error-prone string manipulation
• Automatic validation catches errors before they break reports
• No dependencies - just Python 3.8+ standard library

Installation

Requirements:

• Python 3.8 or higher
• uv (Python package manager and tool runner)

Installing uv:

• macOS/Linux: curl -LsSf https://astral.sh/uv/install.sh | sh
• Windows: powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
• Alternative (all platforms): pip install uv or see installation docs

Note: uvx (included with uv) automatically handles the Python environment and dependencies. No manual Python package installation needed!

Quick Start

{
  "mcpServers": {
    "rdl-mcp": {
      "command": "uvx",
      "args": ["rdl-mcp"]
    }
  }
}

{
  "servers": {
    "rdlMcp": {
      "type": "stdio",
      "command": "uvx",
      "args": ["rdl-mcp"]
    }
  }
}

After installation: Restart your AI assistant and try: "Describe the structure of my report.rdl file"

Usage

Just ask your AI assistant in natural language:

• "What datasets does this report use?"
• "Make the Account Number column 2 inches wide"
• "Format the Amount column as currency with 2 decimals"
• "Add a new Amount column that shows the sum in the footer"
• "Add a Status column but leave the footer blank"
• "Update the main dataset to use the V2 stored procedure and add the TaxAmount field"
• "Remove the obsolete Status column"
• "Add a Year parameter to filter the report"

The AI assistant will use the appropriate MCP tools automatically.

Example: Editing vs. XML

Without MCP (manually editing XML):

<TablixCell><CellContents><Textbox><Paragraphs>
  <Paragraph><TextRuns><TextRun>
    <Value>Old Header</Value>
  </TextRun></TextRuns></Paragraph>
</Paragraphs></Textbox></CellContents></TablixCell>

With MCP (one command):

update_column_header(filepath="report.rdl",
                     old_header="Old Header",
                     new_header="New Header")

API Reference

Limitations & Roadmap

Current limitations:

• Tablix (table) controls only - no Matrix or Chart support yet
• Works best with standard report layouts
• Some complex RDL features may still need manual XML editing

Planned features:

• Column reordering, grouping, and sorting configuration
• Expression builder helpers
• Dataset field management

Troubleshooting

Server not appearing?

• Check absolute path in config is correct
• Verify Python 3.8+: python3 --version
• Restart your MCP client

Permission errors?

• Make script executable: chmod +x rdl_mcp_server.py
• Check RDL file read/write permissions

Releasing a New Version

This server is published to PyPI and the MCP Registry. To release a new version:

1. Update version numbers in both files:

   pyproject.toml:

   version = "0.2.0"
   

   server.json:

   {
     "version": "0.2.0",
     "packages": [
       {
         "version": "0.2.0"
       }
     ]
   }
   

2. Commit your changes:

   git add .
   git commit -m "Release v0.2.0: Add feature description"
   

3. Create and push a git tag:

   git tag v0.2.0
   git push origin main --tags
   

4. Automated publishing: The GitHub Actions workflows automatically:

   • Build and publish to PyPI (users can install via uvx rdl-mcp)
   • Validate server.json against the MCP schema
   • Publish to the MCP Registry (server appears in registry search)
   • Update downstream registries (like GitHub's MCP marketplace)

Contributing

PRs welcome! Priority areas:

• Better column detection for complex layouts
• More editing operations (reordering, grouping, etc.)

Requirements: Python standard library only

1. Fork repo
2. Create feature branch
3. Make changes + tests
4. Submit PR

License

MIT License - see LICENSE file for details.

This means you're free to use, modify, and distribute this software for any purpose, commercial or non-commercial.