HUMANENDPOINT / MCP
Human verification through MCP.
Connect an MCP-compatible agent to HumanEndpoint and give it tools for discovering human-only capabilities, checking task policy, requesting commercial terms, creating jobs, tracking execution and retrieving evidence.
OPEN API SPECVIEW CAPABILITIESMCP endpoint
https://humanendpoint.au/mcpHumanEndpoint is published in the official MCP Registry as au.humanendpoint/humanendpoint.
Useful MCP capabilities
List live human capabilities and fall back to custom-human-task when no named service matches.
Request AI-output verification, workflow QA, research judgement, or evidence-backed physical verification.
Request a quote, accept human-reviewed terms and use the returned x402 payment path.
Poll status, read structured results and verify downloaded artifacts against their SHA-256 hashes.
Machine-readable discovery
OpenAPI, llms.txt, A2A agent card, trust metadata and the MCP interface all point to the same production workflow.
Start with HumanEndpoint
Agents should check /health, read /trust.json, discover capabilities at /v1/services, and use custom-human-task when an unusual lawful outcome does not match a named capability.