LEARNING CENTER

How automated interactions become verifiable

Four guides covering records, verification, policy, and evidence for automated interactions. Each one explains a protocol surface, shows what the wire format looks like, and points to working code.

In-depth guides

31 min

Total read time

Key concepts

When an AI agent calls an API, scrapes a page, or uses a tool, both sides need proof of what occurred. Logs explain what your systems observed; they do not create portable proof another party can verify independently. These guides explain how signed records work, how policy and consent fit in, and how payment evidence layers on top.

Start Learning

Each guide builds on the others. We recommend starting with Verifiable Interaction Records.

START HERE

Verifiable Interaction Records

Why Logs Are Not Enough

What a signed interaction record contains, how offline verification works, and why server logs are not portable proof. Includes the JWS payload structure and Ed25519 verification flow.

Interaction RecordsOffline VerificationPortable Proof

6 min readRead guide

Policy, Consent & Attribution

Machine-Readable Terms for AI

How publishers declare what agents may do with their content (AIPREF, peac.txt) and how agents prove they followed those terms. The gap between robots.txt and what AI actually needs.

AIPREFPolicy DiscoveryContent Terms

7 min readRead guide

HTTP 402 & Payment Evidence

Payment Challenges and Settlement Proof

How HTTP 402 Payment Required works for machine-to-machine interactions. The request/challenge/payment/receipt cycle, x402 protocol headers, and how payment evidence fits into the verification stack.

HTTP 402Payment EvidenceSettlement Proof

10 min readRead guide

Agentic Commerce

The Economic Layer for AI Agents

What happens when machines need to buy API calls, pay for content access, or settle charges with other agents. Covers policy discovery, pricing headers, payment rails, and signed records.

Agent EconomyMachine PaymentsAPI Monetization

8 min readRead guide