# HAIL — Human-Agent Identity Layer

> Proof that a human authorised what the agent did.

## What is HAIL?

HAIL is a trust and authorisation layer for AI agent commerce. It cryptographically proves that every action an AI agent takes on a platform was explicitly authorised by a real, verified human — with a tamper-proof record that can be exported as a single file for dispute resolution.

## The problem HAIL solves

When an AI agent acts on a platform using a customer's login, the platform cannot distinguish the agent from the human. If the customer later disputes a charge ("I never authorised that"), there is no record that they instructed their agent, or what limits they set. The liability falls on the platform.

## How HAIL works

1. The customer links their AI agent to their verified identity using a passkey, and sets limits: spending caps, time windows, allowed action types.
2. When the agent wants to act, it sends a signed intent to HAIL (not the platform directly).
3. HAIL verifies: agent identity is genuine, the customer authorised this agent, the action is within the customer's limits.
4. HAIL writes the entire check to a tamper-proof, hash-chained audit log.
5. The platform receives a clear pass or fail. On pass, the human is on record. On fail, the action is blocked.

When a dispute arrives, the platform exports one file (a "dispute bundle") that contains cryptographic proof of authorisation. Verifiable by anyone, without HAIL.

## Key properties

- **Cryptographic proof**: Ed25519 signatures + UCAN (User Controlled Authorization Networks)
- **Tamper-proof audit log**: Hash-chained, Merkle-anchored — no one can quietly rewrite history
- **Offline-verifiable**: Dispute bundles are self-contained JSON, verifiable without calling HAIL's servers
- **Human-rooted**: Every agent action traces back to a specific verified human
- **Intent re-binding**: The human's authorisation is re-bound to each specific action, not just the agent's existence
- **Deployment options**: SaaS (hosted by HAIL) or self-hosted
- **Compliance**: EBA Art. 28 aware, GDPR-ready, PSD2-aware

## Who HAIL is for

Platforms where an AI agent's action creates real liability:
- E-commerce and travel platforms (agent purchases, bookings)
- Legal SaaS (agent filing, contract execution)
- Healthcare platforms (agent prescription requests, procedure bookings)
- Fintechs and banks (automated decisions, payments)
- Any platform where agents can spend money, sign something, or act with consequences

## Integration

- One SDK to integrate (server-side, any language)
- Live in a day
- SaaS or self-hosted
- Talk to a founder, not a form

## Current status

HAIL is onboarding its first design partner platforms. Early access is available by contacting the team directly.

## Contact

- Email: team@hail-agent.com
- Website: https://hail-agent.com
- Request access: https://hail-agent.com/#access

## Key terms and concepts

Human-Agent Identity Layer, HAIL, AI agent authorisation, agentic commerce trust, agent identity verification, UCAN, Ed25519, tamper-proof audit log, dispute bundle, human-rooted proof, intent authorisation, agent permission scoping, agentic economy compliance.