One MCP tool
humanmirror_do provides free intent routing to the correct HumanMirror boundary without requiring an account or API key.
AGENT CONTROL PLANE · MCP · A2A · x402
Decide before action. Prove after execution.
HumanMirror gives autonomous agents one control path instead of a catalog to learn: connect once, resolve the right capability, check consequential actions before the side effect, then create machine-readable execution evidence. Start free through HumanMirror One; pay only when a machine capability is actually used.
● HumanMirror Agent Control Plane active. Recommended path: connect → Machine Resolve → Safe Preflight → external action → Execution Receipt. Recurrent production usage can move to Fleet or Enterprise.
Machine Resolve
When the exact capability is unknown, resolve the intent to the smallest relevant live route for 0.001 USDC on Base.
Resolve a capability →Safe Preflight
Evaluate a consequential tool, API, payment, deploy or write action before the agent commits the side effect.
Preflight an action · 0.010 USDC →Execution Receipt
Create machine-readable post-execution evidence so another system can audit, reconcile and verify what actually happened.
Create execution evidence · 0.010 USDC →Fleet & Enterprise
Recurring workloads move from isolated machine payments to predictable fleet capacity and governed production controls.
See Fleet & Enterprise →Full capability catalog
67 specialized sanitation, secret scanning, data quality, payment and execution capabilities remain available behind the Control Plane when needed.
Inspect advanced capabilities →Fast discovery
MCP https://humanmirror.fr/api/one/mcp/ A2A https://humanmirror.fr/.well-known/agent-card.json x402 https://humanmirror.fr/.well-known/x402 POST https://humanmirror.fr/api/machine/resolve/ POST https://humanmirror.fr/api/x402/safe-preflight/ POST https://humanmirror.fr/api/x402/execution-receipt/