Agents request work. Humans execute or verify. Proof is structured. Verification gates settlement.
agent request -> human execution / verification -> structured proof -> verify -> settle
I am focused on AI2Human Network, a system for turning human-needed steps into callable infrastructure for agents, protocols, token issuers, and onchain applications.
The core problem is simple:
AI agents can plan and automate software-native workflows, but many real workflows still need verified human execution, evidence, review, and settlement.
AI2Human is building that missing layer.
| Repo | Layer |
|---|---|
ai2human-protocol |
Request manifests, proof specs, verification results, settlement events, task state machines |
ai2human-skills |
Agent-readable skills, manifests, task templates, OpenClaw test paths |
ai2human-proof-kit |
Structured proof schemas, proof bundle examples, verification outputs |
ai2human-settlement-contracts |
Base USDC prize pools, verified reward settlement, payout flows |
ai2human-b20-gateway |
B20 token rules connected with AI2Human human verification |
- AI2Human X Bot: tweet a task, route a human, verify proof, settle onchain
- Agent Skills: allow agents to create campaigns, request verification, and monitor settlement
- B20 Skills: generate B20 token policy plans with human proof requirements
- Proof Kit: make human output machine-readable and reviewable
- Settlement Rails: Base USDC prize pools, escrow, claims, refunds, and payout records
- AI agents and skill runtimes
- Structured proof and verification workflows
- Base / USDC settlement
- B20, RWA, and compliance proof gateways
- Next.js, TypeScript, Solidity, EVM integrations
- AI2Human: ai2human.io
- GitHub org: ai2humannetwork
- X: @ai2humannetwork