SEC MCP

SEC MCP 📊

A Server-Sent Events (SSE) Model Context Protocol server for SEC EDGAR data access. Enables both remote and local connections to retrieve SEC filing data, company information, and financial facts.

Integrations: Claude Desktop • Cursor • Cline • Roo Coder

📋 Table of Contents

• Overview
• Features
• Installation
• Quick Start
• IDE Integrations
• API Examples
• Development
• Troubleshooting

Overview

SEC MCP provides a streamlined interface to access SEC EDGAR data through the Model Context Protocol. It supports real-time data streaming via SSE, making it ideal for both local development and remote deployment scenarios.

Built with Python and the MCP framework, this server enables seamless integration with AI assistants and other tools that support the Model Context Protocol.

Features

• 🔍 Company Search - Search for companies by name, CIK, or ticker
• 📄 Filing Access - Download and retrieve SEC filings
• 📊 Financial Data - Access company facts and concept data
• 🚀 SSE Support - Real-time data streaming for remote connections
• ⚡ Async Operations - High-performance async API client
• 🔒 Rate Limiting - Built-in rate limiting for SEC compliance

Installation

Usage

Run the MCP server using uv:

The server will start on the specified port, ready to accept both local and remote SSE connections.

🚀 Claude Desktop Integration

✨ Quick Setup (3 steps)

1. Install SEC-MCP
   git clone https://github.com/LuisRincon23/SEC-MCP.git cd SEC-MCP uv sync
2. Configure Claude DesktopOpen your Claude Desktop configuration file:
   • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
   • Windows: %APPDATA%\Claude\claude_desktop_config.json

   Add this configuration:

   { "mcpServers": { "SEC-MCP": { "command": "/path/to/SEC-MCP/start-mcp.sh" // macOS/Linux // "command": "C:\\path\\to\\SEC-MCP\\start-mcp.bat" // Windows } } }

   💡 Important: Replace /path/to/SEC-MCP with your actual installation path

3. Restart Claude DesktopThat's it! The SEC EDGAR tools are now available in your conversations.

🎯 Verify Installation

Type any of these in Claude to see the tools in action:

• "Search for Apple's SEC filings"
• "Get Tesla's financial data"
• "Show me Microsoft's recent 10-K"

🛠️ Alternative Setup Methods

If you prefer not to use the shell script, you can configure Claude Desktop directly:

Find your uv path with: which uv

For remote access or advanced setups:

1. Start the server:
   uv run -m run --port 8000
2. Configure Claude Desktop:
   { "mcpServers": { "SEC-MCP": { "url": "http://localhost:8000/sse" } } }

🔧 Troubleshooting

If you encounter issues:

1. Enable Developer Mode (for detailed logs):
   echo '{"allowDevTools": true}' > ~/Library/Application\ Support/Claude/developer_settings.json
   Then use Cmd+Option+Shift+I in Claude Desktop to see logs.
2. Check Logs:
   tail -f ~/Library/Logs/Claude/mcp-server-SEC-MCP.log
3. Common Issues:
   • "spawn ENOENT": The path to uv or the script is incorrect
   • "No module named...": Working directory (cwd) is incorrect
   • Server disconnects: Check that all dependencies are installed with uv sync

🔌 IDE & Code Assistant Integrations

💻 Cursor IDE Integration

Cursor is an AI-powered IDE that supports MCP servers natively.

1. Open Cursor Settings
   • Press Cmd+, (macOS) or Ctrl+, (Windows/Linux)
   • Navigate to Features → MCP Servers
2. Add SEC-MCP Configuration
   { "SEC-MCP": { "command": "/path/to/SEC-MCP/start-mcp.sh" // macOS/Linux // "command": "C:\\path\\to\\SEC-MCP\\start-mcp.bat" // Windows } }
3. Restart Cursor
   • The SEC tools will be available in Cursor's AI assistant
   • Type @SEC-MCP to access the tools directly

Usage Example:

🤖 Cline (VS Code Extension) Integration

Cline is a powerful VS Code extension that brings Claude to your IDE.

1. Install Cline Extension
   • Open VS Code
   • Go to Extensions (Cmd+Shift+X)
   • Search for "Cline" and install
2. Configure MCP Server
   • Open VS Code settings (Cmd+,)
   • Search for "Cline MCP"
   • Add to cline.mcpServers:
   { "SEC-MCP": { "command": "/path/to/SEC-MCP/start-mcp.sh" // macOS/Linux // "command": "C:\\path\\to\\SEC-MCP\\start-mcp.bat" // Windows } }
3. Access SEC Tools
   • Open Cline chat (Cmd+Shift+P → "Cline: Open Chat")
   • SEC tools are now available in your coding sessions

Usage Example:

🦘 Roo Coder Integration

Roo Coder is an AI coding assistant that supports MCP protocol.

1. Install Roo Coder
   npm install -g roo-coder
2. Configure MCP Server Create or edit ~/.roo-coder/config.json:
   { "mcpServers": { "SEC-MCP": { "command": "/path/to/SEC-MCP/start-mcp.sh", // macOS/Linux // "command": "C:\\path\\to\\SEC-MCP\\start-mcp.bat", // Windows "env": { "PYTHONPATH": "/path/to/SEC-MCP" } } } }
3. Launch Roo Coder
   roo-coder --enable-mcp

Usage Example:

📊 Integration Comparison

⚡ Quick Integration Tips

• All IDEs: Ensure the full path to start-mcp.sh is used
• Windows Users: Use forward slashes in paths or escape backslashes
• Permissions: Make sure start-mcp.sh is executable (chmod +x start-mcp.sh)
• Python Environment: The shell script handles uv environment automatically
• Testing: Use the verification examples in each section to test your setup

Available Tools

1. search_companies - Search for companies by name, CIK, or ticker
2. get_company_submissions - Retrieve all SEC submissions for a company
3. get_company_facts - Get standardized company facts data
4. get_company_concept - Access specific XBRL concepts for a company
5. download_filing - Download filing documents by URL

Example Client Connection

API Examples

Search Companies

Get Company Filings

Access Financial Data

Configuration

The server accepts the following command-line arguments:

• --port - Port number for the SSE server (default: 8000)
• --host - Host address to bind to (default: localhost)
• --transport - Transport type: stdio or sse (default: sse)

Requirements

• Python 3.9+
• uv package manager
• Dependencies managed via pyproject.toml

Development

License

MIT License - see LICENSE file for details

Author

Created by Luis Angel Rincon

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Acknowledgments

Built using the Model Context Protocol (MCP) and SEC EDGAR API.