MCP Integration

Use Coda from Claude Desktop, Cursor, and other MCP clients.

Coda exposes its quantum computing tools as an MCP (Model Context Protocol) server. This lets you write and run quantum algorithms, simulate circuits, transpile between frameworks, and submit to real quantum hardware - all directly from AI-powered editors and assistants.

Supported MCP clients

Cursor
Claude Desktop
Claude Code
Codex
Any client that supports the MCP standard

Setup

Add the Coda MCP server to your client configuration. Select your client below to see the setup instructions.

1 {
2   "mcpServers": {
3     "coda": {
4       "command": "/path/to/uvx",
5       "args": ["coda-mcp"],
6       "env": {
7         "CODA_API_TOKEN": "<your-api-key>"
8       }
9     }
10   }
11 }

Find your uvx path by running which uvx in your terminal.

Client	Config location
Cursor	`~/.cursor/mcp.json`
Claude Desktop	`~/Library/Application Support/Claude/claude_desktop_config.json` (macOS)
Claude Code	`~/.claude.json`, or run the command above in your terminal
Codex	`~/.codex/config.toml` under `[mcp_servers.coda]`, or run the command above

Available tools

Once connected, the MCP server exposes the following tools:

Tool	Description
`simulate`	Simulate a quantum circuit
`transpile`	Convert between quantum frameworks
`to_openqasm3`	Convert to OpenQASM 3.0
`estimate_resources`	Analyze circuit complexity
`split_circuit`	Cut large circuits
`qpu_submit`	Submit to quantum hardware
`qpu_status`	Check job status
`qpu_devices`	List available devices

Next steps

Coda API Reference: full REST API documentation
Quickstart: get started with the REST API

1	{
2	"mcpServers": {
3	"coda": {
4	"command": "/path/to/uvx",
5	"args": ["coda-mcp"],
6	"env": {
7	"CODA_API_TOKEN": "<your-api-key>"
8	}
9	}
10	}
11	}