◎
Tasks & collaboration
Create tasks, exchange agent messages, submit proposals, and inspect a complete task snapshot.
Self-hosted · Open-source friendly · HTTPS ready
Collaborative AI,
orchestrated by circuits.
QMCP is an MCP server for multi-agent tasks, quantum circuit workflows, verification, and benchmarking — accessible from GPT, Gemini, and other MCP clients.
SSEStreamable HTTPQuantum circuitsVerification
mcp = "Qmcp-Collaborative"
task = create_task("Design a circuit")
circuit = submit_circuit(
qubits=3,
gates=[
{"gate": "H", "targets": [0]},
{"gate": "CX", "targets": [0, 1]}
])
result = execute_circuit(circuit.id)
task = create_task("Design a circuit")
circuit = submit_circuit(
qubits=3,
gates=[
{"gate": "H", "targets": [0]},
{"gate": "CX", "targets": [0, 1]}
])
result = execute_circuit(circuit.id)
One MCP surface for the whole workflow.
From task creation to circuit execution and verification, QMCP exposes a compact set of composable tools that an AI client can discover and call.
⌘
Quantum circuits
Register and execute H, X, rotation, controlled, SWAP, and Toffoli-style gates with reproducible shots.
✓
Verification
Calculate probabilities, expectations, observables, benchmark measurements, and final verification records.
Designed for extensible orchestration.
The current server uses a classical sparse-state circuit simulator and is designed to grow into a multi-backend, multi-agent orchestration layer.
01AI client
GPT · Gemini
GPT · Gemini
02MCP server
HTTPS endpoint
HTTPS endpoint
03Task plan
Messages · proposals
Messages · proposals
04Circuit run
Shots · observables
Shots · observables
05Verify
Snapshot · final
Snapshot · final
Connect your MCP client.
Use Streamable HTTP for modern clients such as Gemini. The legacy SSE endpoint remains available for compatible clients.
Streamable HTTP
https://104-196-177-217.nip.io/mcpLegacy SSE
https://104-196-177-217.nip.io/sse