Use as MCP server

The calculator on this site is also a Model Context Protocol server. Add the URL once to your AI assistant's config and the compute tool, the historic hardware data, and the curated algorithm list become native tools — no API key, no bootstrap fetch, no JSON-parsing boilerplate.

Endpoint

https://www.quantum-expectations.com/api/mcp

Streamable HTTP. Stateless sessions. CORS open. Rate-limited per IP.

Claude Desktop / Cursor / Windsurf

Add to your MCP config file:

{
  "mcpServers": {
    "quantum-expectations": {
      "type": "http",
      "url": "https://www.quantum-expectations.com/api/mcp"
    }
  }
}

Claude Desktop: ~/Library/Application Support/Claude/claude_desktop_config.json on macOS. Restart the client after saving.

Claude Code CLI

One command:

claude mcp add --transport http quantum-expectations https://www.quantum-expectations.com/api/mcp

ChatGPT

Settings → Connectors → Add custom connector, then paste the endpoint URL above.

Stdio-only clients

Bridge through mcp-remote:

{
  "command": "npx",
  "args": ["-y", "mcp-remote", "https://www.quantum-expectations.com/api/mcp"]
}

What you get

compute_expectation — effective error rate, success probability, surface-code / qLDPC overhead for a given (p, n, d).
list_current_quantum_computers, list_example_algorithms, list_qldpc_codes — reference data (same rows you see on this site).
get_historic_fidelity_series, get_historic_qubit_count_series — full time series with source URLs for trend extrapolation.
get_agent_brief — scope, assumptions, honesty clause.

Try it

Once connected, ask your assistant:

"Using the quantum-expectations MCP server, compute the effective error rate for p=1e-3, n=100, d=1000 without error correction, and explain the result."

Back Full agent brief (/agent.txt)