A Model Context Protocol (MCP) proxy server that bridges AI assistants (Claude Desktop, Gemini CLI, etc ...) to a remote Eulerian Marketing Platform MCP server. This proxy handles authentication, request forwarding, and provides a local MCP interface to your remote Eulerian instance.
How It Works
Provides access to the Eulerian Marketing Platform: analytics, attribution & data activation.
API key
EMP_API_TOKEN, EMP_API_ENDPOINT
ตรวจสอบล่าสุด
2026-11-09
ดาว
—
เครื่องมือ
—
อัปไทม์
—
This server acts as a transparent proxy between local MCP clients and your remote Eulerian Marketing Platform server:
┌─────────────┐ ┌──────────────────┐ ┌────────────────┐
│ Claude │ ◄─────►│ This MCP Proxy │◄─────► │ Remote Eulerian│
│ Desktop │ stdio │ (Local) │ HTTP │ MCP Server │
└─────────────┘ └──────────────────┘ └────────────────┘
The proxy:
🔐 Handles authentication with your remote Eulerian server
📡 Forwards MCP requests via HTTP with Bearer token
🛠️ Exposes remote tools and resources to AI assistants
📝 Provides comprehensive logging for debugging
⚡ Uses async HTTP for better performance
Warning
Given the rapid pace of evolution of protocols, implementations and new features coming out, the MCP may or may not be compatible with your suite of tools.
Please fill in an issue so that we can check it.
Features
🔌 Proxy Architecture: Bridges local MCP clients to remote Eulerian MCP server via HTTP
🔐 Secure Authentication: Uses Bearer token authentication for remote server access
🌐 Cross-platform support: Works on Windows, Linux, and macOS
🤖 Multiple AI clients: Compatible with Claude Desktop and Gemini CLI
📝 Comprehensive Logging: Logs all requests/responses for debugging
⚡ Async HTTP: Non-blocking requests using httpx for better performance
🛠️ Tool Discovery: Automatically discovers and exposes remote tools
Note: The section name must use mcp_servers (with an underscore). Using mcp-servers or any other variant will silently fail.
Option C: Project-scoped configuration
To limit the MCP server to a specific project, create a .codex/config.toml file at the root of that project with the same content as above. The project must be marked as trusted by Codex.
Verify the connection
Launch Codex in your terminal:
codex
Type /mcp in the interactive TUI to see all connected MCP servers.
Confirm that eulerian-marketing-platform appears in the list with its available tools.
Try asking:
What Eulerian Marketing Platform tools do you have access to?
Managing the server
# List all configured MCP servers
codex mcp
# Remove the server
codex mcp remove eulerian-marketing-platform
Troubleshooting (Codex CLI)
Server not appearing in /mcp: Verify that the [mcp_servers.eulerian-marketing-platform] section is present in ~/.codex/config.toml and that the TOML syntax is valid.
Timeout errors: Increase tool_timeout_sec in config.toml (default is 60 seconds). Eulerian queries can take longer, so 300 is recommended.
Authentication errors: Double-check that EMP_API_ENDPOINT and EMP_API_TOKEN are correct.
Python not found: Ensure the python command resolves to Python 3.10+. You may need to use python3 instead:
command = "python3"
Package not found: Make sure eulerian-marketing-platform is installed in the Python environment that Codex will invoke. Run python -m eulerian_marketing_platform.server manually to confirm it works.
Check logs: Monitor the proxy logs for detailed error information:
tail -f /tmp/eulerian-mcp-proxy.log
4. Claude Code
Claude Code is Anthropic's terminal-based coding agent. It supports MCP servers configured via the claude mcp CLI command or by editing ~/.claude.json.
Prerequisites
Node.js 18+ and npm
Claude Code installed: npm install -g @anthropic-ai/claude-code
Python 3.10+ with this package installed (pip install eulerian-marketing-platform)
An Anthropic API key or Claude subscription
Option A: Using the claude mcp add command (easiest)
Type /mcp to see the status of all connected MCP servers.
Confirm that eulerian-marketing-platform appears and shows as connected.
Try asking:
What Eulerian Marketing Platform tools do you have access to?
Managing the server
# List all configured MCP servers
claude mcp list
# Get details for a specific server
claude mcp get eulerian-marketing-platform
# Remove the server
claude mcp remove eulerian-marketing-platform
Troubleshooting (Claude Code)
Server not appearing in /mcp: Run claude mcp list to confirm it is registered. Check ~/.claude.json for JSON syntax errors.
Python not found: Ensure the python command resolves to Python 3.10+. You may need to use python3 instead.
Package not found: Make sure eulerian-marketing-platform is installed in the Python environment that Claude Code will invoke.
Debug mode: Launch Claude Code with verbose MCP logging:
claude --mcp-debug
Check logs: Monitor the proxy logs for detailed error information:
tail -f /tmp/eulerian-mcp-proxy.log
5. Cursor
Cursor is an AI-powered code editor with built-in MCP support.
Alternatively, to make the server available globally across all workspaces, add the configuration to your User Settings (JSON) (Ctrl+Shift+P → Preferences: Open User Settings (JSON)):
VS Code will prompt you for the token when the server starts.
Troubleshooting (VS Code)
MCP not available: Ensure you are running VS Code 1.101 or later and have the GitHub Copilot extension installed and enabled.
Tools not appearing: MCP tools only work in Agent mode. Toggle to Agent mode in the Copilot Chat panel.
Start button not showing: The code lens (Start button) only appears if the .vscode folder is at the root of your workspace.
Python not found: Ensure the python command resolves to Python 3.10+. You may need to use python3 or a full path.
Check logs: Monitor the proxy logs:
tail -f /tmp/eulerian-mcp-proxy.log
Usage Examples
Once configured with any client, you can interact with your remote Eulerian Marketing Platform:
User: "What tools are available from Eulerian?"
→ Proxy calls list_remote_tools() and returns all available tools
User: "Call the update_goal tool"
→ Proxy forwards to remote server and update goal settings
User: "Show me campaign details for CAMP-12345"
→ Claude uses call_eulerian_tool() to fetch specific campaign
User: "What resources are available?"
→ Proxy lists all available data sources
The AI assistant will automatically use the appropriate proxy tools to fulfill your requests.
"EMP_API_ENDPOINT environment variable is required"
Solution: Ensure you've set the EMP_API_ENDPOINT in your configuration
Check that there are no typos in the environment variable name
"EMP_API_TOKEN environment variable is required"
Solution: Ensure you've set the EMP_API_TOKEN in your configuration
Verify your token is valid and hasn't expired
Server not appearing in Claude Desktop
Solution:
Restart Claude Desktop completely
Check the configuration file for JSON syntax errors
Verify the file path in your config is correct
Look at logs:
macOS: ~/Library/Logs/Claude/mcp-server-*.log
Windows: %APPDATA%\Claude\logs\
Tools not showing in Gemini CLI
Solution:
Use /mcp command to check server status
Verify the settings.json is valid JSON
Restart Gemini CLI
Debug Mode
For detailed debugging:
# Run with MCP Inspector
npx @modelcontextprotocol/inspector uvx eulerian-marketing-platform
# Or with environment variables
EMP_API_ENDPOINT=your_endpoint EMP_API_TOKEN=your_token uvx eulerian-marketing-platform
Development
Running Tests
# Install development dependencies
pip install -e ".[dev]"
# Run tests
pytest
# Run with coverage
pytest --cov=eulerian_marketing_platform
Building from Source
# Clone the repository
git clone https://github.com/EulerianTechnologies/eulerian-marketing-platform-mcp.git
cd eulerian-marketing-platform-mcp
# Install in development mode
pip install -e .
# Build Python distribution (wheel + tarball)
pip install build
python -m build
# Build DXT extension for Claude Desktop
npx @anthropic-ai/dxt pack .
# → eulerian-marketing-platform-mcp.dxt
Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
Fork the repository
Create your feature branch (git checkout -b feature/amazing-feature)
Commit your changes (git commit -m 'Add amazing feature')
Push to the branch (git push origin feature/amazing-feature)
Open a Pull Request
License
This project is licensed under the MIT License - see the LICENSE file for details.