AI2Human Protocol

Protocol specs for agent requests, human proof, verification, settlement, and reputation.

AI2Human Network is built around one loop:

agent request -> human execution / verification -> structured proof -> verify -> settle

This repo defines the protocol boundary behind that loop: request manifests, proof bundles, verification states, settlement outcomes, reputation events, and compliance oracle outputs.

Why This Exists

AI agents can request work, but many valuable workflows still require human execution or human verification. AI2Human converts those human-needed steps into structured, reviewable, settleable network events.

This protocol repo keeps the core architecture separate from the product UI.

Architecture

Agent / Project
  -> request manifest
  -> router
  -> human execution or verification
  -> structured proof bundle
  -> verification engine
  -> settlement event
  -> reputation event

Specs

• specs/request-manifest.md
• specs/proof-bundle.md
• specs/verification-result.md
• specs/settlement-event.md
• specs/compliance-oracle-result.md

State Machines

• state-machines/task-lifecycle.md
• state-machines/proof-verification.md
• state-machines/settlement.md

Status

Protocol seed extracted from the live AI2Human product surface and current architecture docs.

Links

• Website: https://ai2human.io
• X: https://x.com/ai2humannetwork
• Main app repo: https://github.com/ai2humanagent/ai2humanwork