coda-mcp 0.2.2

MCP server for quantum computing tools - simulation, transpilation, QPU access

coda-mcp

MCP server for quantum computing tools. Provides circuit simulation, transpilation, and QPU access.

Works with MCP-compatible clients that support stdio transport: Claude Desktop, Claude Code (CLI), VS Code, Cursor, Zed, and more.

Quick Start

1. Get your API token

Generate a token at coda.conductorquantum.com/settings/api

2. Configure your MCP client

Claude Desktop

Add to your Claude Desktop config:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %AppData%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "coda": {
      "command": "/path/to/uvx",
      "args": ["coda-mcp"],
      "env": {
        "CODA_API_TOKEN": "your-token-here"
      }
    }
  }
}

Find your uvx path with which uvx (typically ~/.local/bin/uvx or /usr/local/bin/uvx).

Restart Claude Desktop after saving.

Claude Code (CLI)

Quick install via CLI:

claude mcp add --env CODA_API_TOKEN=your-token-here coda -- uvx coda-mcp

Or manual configuration - add to ~/.claude.json:

{
  "mcpServers": {
    "coda": {
      "command": "/path/to/uvx",
      "args": ["coda-mcp"],
      "env": {
        "CODA_API_TOKEN": "your-token-here"
      }
    }
  }
}

Team-shared config - add .mcp.json to your project root:

{
  "mcpServers": {
    "coda": {
      "command": "/path/to/uvx",
      "args": ["coda-mcp"],
      "env": {
        "CODA_API_TOKEN": "${CODA_API_TOKEN}"
      }
    }
  }
}

Then team members only need to set CODA_API_TOKEN in their environment.

Cursor

Add to your Cursor MCP config (~/.cursor/mcp.json or .cursor/mcp.json in project root):

{
  "mcpServers": {
    "coda": {
      "command": "/path/to/uvx",
      "args": ["coda-mcp"],
      "env": {
        "CODA_API_TOKEN": "your-token-here"
      }
    }
  }
}

Find your uvx path with which uvx.

Zed

Add to your Zed settings (~/.config/zed/settings.json):

{
  "context_servers": {
    "coda": {
      "command": {
        "path": "/path/to/uvx",
        "args": ["coda-mcp"],
        "env": {
          "CODA_API_TOKEN": "your-token-here"
        }
      }
    }
  }
}

Generic MCP Client

For any MCP-compatible client, configure:

• Transport: stdio
• Command: uvx coda-mcp
• Environment: CODA_API_TOKEN=your-token-here

Or if you prefer pip installation:

pip install coda-mcp
coda-mcp  # runs the server

Installation

# Using uvx (recommended, no install needed)
uvx coda-mcp

# Using pip
pip install coda-mcp

# Using uv
uv tool install coda-mcp

Tools

Quantum Circuit Tools

Tool	Description
transpile	Convert between quantum frameworks (Qiskit, Cirq, PennyLane, Braket, PyQuil, CUDA-Q, OpenQASM)
simulate	Run circuit simulation (CPU via Aer, GPU via CUDA-Q)
to_openqasm3	Convert circuit to OpenQASM 3.0
estimate_resources	Analyze qubit count, depth, and gate counts
split_circuit	Cut large circuits for distributed execution

QPU Tools

Tool	Description
qpu_submit	Submit circuit to QPU backend (IonQ, IQM, Rigetti, AQT, IBM Quantum)
qpu_status	Check job status and get results
qpu_devices	List available QPU devices

Examples

Simulate a Bell State

Create and simulate a Bell state circuit

The simulate tool accepts Qiskit code:

from qiskit import QuantumCircuit
qc = QuantumCircuit(2)
qc.h(0)
qc.cx(0, 1)
qc.measure_all()

Submit to Real QPU

Submit my Qiskit circuit to IBM Torino

Use qpu_submit with code="..." plus source_framework="qiskit" (or another supported framework).

Then call qpu_status with the returned job ID.

Environment Variables

Variable	Description	Required
CODA_API_TOKEN	Your API token from coda.conductorquantum.com/settings/api	Yes
CODA_API_URL	API endpoint URL (default: production API)	No

Troubleshooting

"CODA_API_TOKEN not set"

• Ensure the environment variable is set in your MCP client config
• Verify your token is valid at coda.conductorquantum.com/settings/api

"Connection refused"

• Check your internet connection
• The Coda API may be temporarily unavailable

Tools not appearing

• Restart your MCP client after configuration changes
• Verify uvx coda-mcp runs without errors in terminal
• Check that JSON config syntax is valid

Debug logs (Claude Desktop)

• macOS: ~/Library/Logs/Claude/mcp.log and mcp-server-coda.log
• Windows: %AppData%\Claude\logs\

Debug logs (Claude Code)

claude mcp list        # Check server status
/mcp                   # Check status in Claude Code session

Windows-specific issues

• If uvx is not found, use full path or install via pip install coda-mcp
• Add APPDATA to env if you see ENOENT errors

License

MIT