x402
Open, HTTP-native payment standard from Coinbase that revives the 402 status code so agents can pay for APIs and content per request.
Official site ↗Directory · Research · Protocols
AI agents are starting to buy, subscribe and pay on our behalf. We track the tools, protocols and practices that keep those transactions authorised, limited and fraud-resistant.
01 — Directory
Real products and standards for agent wallets, payment authorization and spend control. Listings are editorial and unpaid.
Open, HTTP-native payment standard from Coinbase that revives the 402 status code so agents can pay for APIs and content per request.
Official site ↗Google-led open protocol using cryptographically signed mandates to prove user intent when agents make purchases.
Official site ↗Toolkit for giving AI agents onchain wallets with configurable policies, built on Coinbase Developer Platform.
Official site ↗SDK for integrating Stripe into agent frameworks, including scoped restricted API keys to limit what an agent can do.
Official site ↗Visa's program for agent-initiated payments using tokenized credentials, user-set controls and authentication.
Official site ↗Framework introducing Agentic Tokens so merchants and issuers can recognise and trust AI-agent transactions.
Official site ↗Payment and identity network that lets AI agents hold funds, verify identity and transact with services.
Official site ↗Platform for letting AI agents send money with human-approved policies, spending rules and oversight.
Official site ↗Wallet and payments infrastructure with APIs for agents to hold assets and complete purchases.
Official site ↗Secure-enclave key infrastructure with a policy engine for programmatic, non-custodial wallet signing.
Official site ↗Embedded wallet infrastructure with server wallets and policy controls usable by automated agents.
Official site ↗Card-issuing platform with virtual cards and real-time authorization controls for spend limits and merchant locks.
Official site ↗Smart-account infrastructure used to run agent wallets as programmable accounts with recovery and spending controls.
Official site ↗Smart accounts with session keys and gas sponsorship, letting an agent act inside a bounded, time-limited session.
Official site ↗Enterprise custody and stablecoin infrastructure with a policy engine that approves or blocks transfers before signing.
Official site ↗Blockchain intelligence platform for spotting fraud, sanctions exposure and unusual flows around agent payments.
Official site ↗Screening and audit tooling that checks crypto payments against sanctions lists and known fraud networks.
Official site ↗Explorer for the x402 ecosystem: transactions, sellers, origins and resources agents are actually paying for.
Official site ↗Security, risk and compliance advisor that audits smart contracts and the systems that sign payments.
Official site ↗02 — Standards
Two open standards are shaping how agents pay: one for machine-to-machine micropayments, one for proving a human authorised the purchase.
Uses the long-reserved HTTP 402 Payment Required status. A server responds with payment requirements; the agent signs a payment (typically a stablecoin transfer) and retries the request with a payment header. A facilitator verifies and settles.
The Agent Payments Protocol extends agent-to-agent communication with verifiable digital credentials called mandates. An Intent Mandate captures what the user authorised; a Cart Mandate locks the exact items and price before payment.
03 — Research
Prompt injection, spoofed merchants and manipulated product pages can redirect an agent's payment. Treat every external input as untrusted and require confirmation for new payees.
6 min read · Read guide →
Compare custodial, MPC and secure-enclave wallets. Keep signing keys out of the model's context and let a policy engine, not the LLM, decide what gets signed.
8 min read · Read guide →
Layer per-transaction caps, daily budgets, merchant allowlists and velocity checks. Limits should be enforced at the payment rail, where the agent cannot rewrite them.
5 min read · Read guide →
Scoped tokens, signed mandates and step-up authentication give merchants and issuers evidence of user consent, and give you a defensible trail when disputes arise.
7 min read · Read guide →
A server answers with HTTP 402, the agent signs a small stablecoin payment and retries. Here is the full flow, what the facilitator does, and where the risk sits.
7 min read · Read guide →
Intent, cart and payment mandates are signed credentials. They let a merchant prove the charge matched what the user authorised, and give you evidence in a dispute.
6 min read · Read guide →
Virtual cards with per-merchant locks, spend caps and short expiry keep a compromised agent away from your main account and your real card number.
6 min read · Read guide →
Fifteen checks across keys, limits, merchant allowlists, confirmation flows, logging and incident response, before an agent is allowed to pay on its own.
9 min read · Read guide →