Agentspay

Explainer

x402 Protocol: How HTTP 402 Payment Required Works for AI Agent Payments

x402 took the one HTTP status code the web never used and turned it into a payment rail machines can drive. The protocol is elegant and genuinely small. The part it deliberately leaves to you is the budget.

Agent Payments Console

Pick an agent

Payment intent

intent:

Policy evaluation

Human approval required

This spend is over your approval threshold. Approve it to issue a scoped card, or deny it.

Scoped virtual card issued

Agentspay

single-use

Wallet budget

spent of

Audit trail

In short

x402 is an open payment protocol that lets a client pay for a web resource inside a normal HTTP request, using the long-dormant 402 Payment Required status code. A server protecting a paid endpoint answers an unpaid request with HTTP 402 and a JSON body listing what it will accept: the price, the asset, the network and the address to pay. The client signs a payment, retries the same request with the payment attached in a header, and a third party called a facilitator verifies the signature and settles the transfer on chain. The server then returns the resource. It was created at Coinbase, settles mainly in stablecoins such as USDC across EVM chains and Solana, and moved to open governance under the Linux Foundation, whose x402 Foundation launched operationally on July 14, 2026 with 40 member organizations including Visa, Mastercard, American Express, Google, AWS, Stripe, Adyen and Shopify. One thing x402 does not define is a buyer-side budget. The protocol caps what a seller may charge for a single call, never what your agent may spend across a thousand of them.

What is x402?

x402 is a way for software to pay software over HTTP, with no account, no subscription and no checkout page. The name comes from the status code it revives: 402 Payment Required, reserved in the original HTTP specification and left unimplemented for roughly thirty years because nobody had a payment method a machine could complete on its own. Stablecoins gave it one. The official description on x402.org calls it an "open, neutral standard for internet-native payments" that "absolves the Internet's original sin by natively making payments possible between clients and servers." In practice it means an AI agent can hit a paid API it has never seen before, read the price out of the error response, pay it, and get the data back, all inside a couple of round trips. The protocol is deliberately thin. It describes how to advertise a price and how to carry proof of payment. It does not run your business logic, hold your money, or decide whether a purchase was a good idea.

What is HTTP 402 Payment Required?

402 Payment Required is a client error status code in the 4xx family, set aside in the HTTP standard for exactly this purpose and then never standardized further. For decades the specification effectively said the code was reserved for future use, so servers that wanted to charge for something redirected you to a billing page or returned 401 or 403 instead. That is why the 402 status code is the one most developers have read about but never actually received. If you are hitting a 402 today from an ordinary API, it usually means an account is out of credit or a payment method failed, and each vendor invented its own meaning. x402 is the first serious attempt to give the code a shared, machine-readable contract: a 402 response is no longer a dead end but a quote, structured well enough that a program can act on it without a human reading the docs.

How does x402 work, step by step?

Coinbase documents the flow in nine steps, and every implementation follows the same shape. One, the client requests a protected resource with an ordinary HTTP request. Two, the server answers 402 with a JSON body describing what it will accept: price, recipient, network and asset. Three, the client picks one of the offered options, signs a payment authorization, and resends the identical request with the signed payload in a header. Four, the server passes that payload to a facilitator and asks it to verify. Five, the facilitator confirms the signature and that the funds are good. Six, the server does the actual work, generating the response the client paid for. Seven, the server asks the facilitator to settle. Eight, the facilitator broadcasts and confirms the transfer on chain. Nine, the server returns the resource together with a settlement receipt header. The ordering in steps six and seven is the detail worth noticing: the work happens before settlement, so a seller carries a small amount of risk per call in exchange for not making the buyer wait on a block confirmation.

What is inside a 402 response?

The response body is the whole interface, and it is small enough to read in one sitting. At the top level sits x402Version, a number, and accepts, a non-empty array where each entry is one payment option the server will honor. A client that understands none of the offered options simply gives up; a client that understands several picks whichever it can settle most cheaply. Each entry names a scheme, which says how the payment is constructed. The scheme in production use is exact, meaning a fixed amount transferred to a fixed address. When the scheme is exact, the entry also carries network, maxAmountRequired, resource, payTo, asset and maxTimeoutSeconds. The table below is what each of those actually controls, because one of them is routinely misread as a spending limit and is not one.

What is an x402 facilitator?

A facilitator is the service that stands between an HTTP server and a blockchain so the server never has to speak to a chain directly. It does two jobs and only two: verify, meaning check that a signed payment payload is valid, well formed and backed by funds, and settle, meaning actually move the money on the seller's behalf across whichever networks and assets it supports. That separation is why a Python API can start charging per call without anyone on the team learning to manage wallets, gas or reorgs. Coinbase runs a facilitator through its Developer Platform, and the v2 architecture lets others plug in as first-class citizens rather than forks. Whether you use a hosted one or stand up your own is a real engineering decision with cost, latency and trust consequences, and we walk through it in what an x402 facilitator does and whether to run your own.

What changed in x402 v2?

Version 2 landed on June 24, 2026 and is the version to build against. The headline change is cosmetic but breaking: the protocol dropped the deprecated X- prefixed headers, so X-PAYMENT became PAYMENT-SIGNATURE and X-PAYMENT-RESPONSE became PAYMENT-RESPONSE. The official SDKs detect both and stay backward compatible with v1, which is generous, but anything of yours that inspects headers directly will not: log pipelines, WAF rules, API gateway policies and analytics that filter on the old names go quietly blind rather than failing loudly. Check those before you assume the migration was free. Underneath the rename, v2 is the more consequential release. Chains are now identified with CAIP identifiers instead of bespoke strings, the SDK became plugin-driven so developers register networks, assets and payment schemes rather than patching internals, payTo can be routed per request to different addresses or callback logic, and a Discovery extension lets facilitators index paid endpoints and their prices automatically so agents can find them.

Who controls x402 now?

x402 started at Coinbase and no longer belongs to it. The Linux Foundation announced its intent to host the protocol in April 2026, choosing the second of the month as a nod to the status code, and the x402 Foundation launched operationally on July 14, 2026 with 40 member organizations. The membership is the part worth reading closely, because it is not a crypto-only roster. Premier members include Visa, Mastercard, American Express, Fiserv, Adyen, Stripe, Google, AWS, Cloudflare, Shopify, Circle, Coinbase, Ripple, MoonPay, and the Solana, Stellar and Monad foundations. General and associate members add Fireblocks, Polygon Labs, NEAR, LayerZero, Quant Network, KakaoPay, zerohash and the Cardano Foundation among others. When the card networks and the largest acquirers join the governing body of a stablecoin-settled HTTP protocol, the reasonable read is that they expect machine-to-machine payments to be a separate lane from card checkout rather than a threat to it.

What is x402 actually good for?

The honest scope is narrower than the coverage suggests, and knowing the edges saves you a wasted quarter. x402 is strong where payments are small, frequent and machine-to-machine: paying per API call, per inference, per MCP tool invocation, per scraped page, per dataset query, or agent to agent for a subtask. At those sizes card rails are economically impossible, since a fixed authorization fee dwarfs a payment worth a fraction of a cent, and x402 settles for a fraction of that. It is also unusually good at first contact, because an agent can transact with a service it has no relationship with, no account on and no API key for. Where it does not fit is ordinary retail. Settlement is a stablecoin transfer on chain, so the merchant has to accept that asset, which almost no US storefront does. There is no chargeback, no dispute process and no card network arbitration, so a payment that buys nothing is simply gone. If your agent needs to buy from businesses that take cards, you want a scoped virtual card or a network agentic token instead, which we lay out in agentic payments and the rails agents pay over.

x402 compared to the checkout protocols

It is easy to lump x402 in with the Agentic Commerce Protocol, Google AP2 and UCP because they all arrived during the same eighteen months and all mention agents. They solve different problems. ACP and UCP standardize the handshake between an AI shopping surface and a merchant, so an agent can browse a catalog, build a cart and complete a checkout against an existing payment method. AP2 standardizes the mandate, a signed record of what a human authorized an agent to do. x402 skips shopping entirely: there is no cart, no catalog and no merchant of record, just a resource with a price. The practical consequence is that they compose rather than compete. An agent might use ACP to buy a physical product and x402 to pay for the data it used to decide what to buy, in the same task. The full three-way breakdown is in AP2 vs ACP vs x402.

x402 compared to MPP, the other HTTP 402 standard

Since March 2026 x402 has a direct rival using the same status code. The Machine Payments Protocol, co-authored by Stripe and Tempo and submitted to the IETF standards track, also answers an unpaid request with HTTP 402 and a challenge. Two differences decide which one fits. Payment methods: x402 is blockchain-only, so the payment is always a stablecoin transfer, while MPP treats the method as an extension point and ships production support for stablecoins, cards and Lightning. For a buyer whose treasury will not hold crypto, that is not a preference, it is the whole decision. Production primitives: idempotency, expiration and request-body binding sit in the MPP base specification, whereas around x402 you generally build them yourself. x402 keeps two real advantages: it is permissionless in a way a chain with a curated validator set is not, and it has been running in the open for longer. Neither standard defines a spending limit, which is the subject of the next section.

The gap x402 leaves you: nobody set the buyer a budget

Read the specification looking for a spending limit and you will find maxAmountRequired and think you have found one. You have not. That field is the seller declaring the most it will charge for this one call, and it protects the seller from a client that underpays, not you from an agent that overspends. Nothing in x402 knows how much your agent has already spent this hour, whether a counterparty is one you trust, or whether a person would have approved this if asked. The protocol is stateless by design and that is a reasonable design; budgets are a property of the payer, not of the payment. It does mean that if you point an agent at a wallet and let it pay 402s, the ceiling on your losses is the wallet balance, and a retry loop against a mispriced endpoint will find that ceiling faster than your alerting will. The control has to live above the rail: a hard per-agent budget, a per-counterparty cap, an approval threshold above which a human is asked, and a record that ties every settled payment back to the agent and the task that caused it. AWS built exactly this shape into Bedrock AgentCore Payments, where a payment session carries a maxSpendAmount and the limit check runs before the payload is signed, so a denied payment never touches the chain. We cover the ceiling question directly in can you set spending limits on x402, and Agentspay enforces those limits across x402 and card rails alike from one place.

The clearest illustration now ships from AWS. Amazon Bedrock AgentCore Payments reached general availability on 18 August 2026 as a managed x402 client, and it does set a ceiling, which is more than most implementations bother with. The ceiling is one optional amount per payment session, denominated only in USD, on a session that expires after at most 480 minutes, with no cumulative total anywhere in its API. We measured the whole surface on our AWS AgentCore spend controls page.

Whatever standard moves the money, Agentspay is the rail-neutral control plane that keeps it governed. See how it works and the control surfaces that enforce policy, approvals, and audit on every transaction.

The 402 response

Every field in an x402 payment requirement, and what it controls

Fields carried by each entry in the accepts array when the scheme is exact. Note that none of them is a buyer-side budget.

Field What it is Who sets it Common mistake
x402Version Protocol version at the top level of the response, a number Server Assuming a v1 client can read a v2 response without SDK support
accepts Array of payment options the server will honor, at least one Server Handling only the first entry instead of picking the cheapest settleable one
scheme How the payment is constructed. In production this is exact Server Building for schemes that are not yet deployed
network Which chain settles the payment, CAIP identified since v2 Server Hardcoding a chain string that v2 renamed
maxAmountRequired The most the seller will charge for this single call Server Reading it as a spending cap for your agent. It is not
payTo Destination address for the funds, routable per request in v2 Server Caching it, when v2 allows it to change per call
asset Token contract being paid, typically USDC Server Assuming every facilitator supports every asset
resource The URL the payment buys, binding payment to endpoint Server Replaying a payload against a different endpoint
maxTimeoutSeconds How long the server will wait for settlement to complete Server Setting client timeouts shorter than this value

Version change

x402 v1 vs v2: what actually changed

v2 shipped June 24, 2026. The SDKs accept both, but anything reading headers outside the SDK does not.

x402 v1 x402 v2
Request header X-PAYMENT PAYMENT-SIGNATURE
Response header X-PAYMENT-RESPONSE PAYMENT-RESPONSE
Chain identifiers Bespoke network strings CAIP standard identifiers
Adding a chain or asset Patch the SDK internals Register a plugin, no fork required
Payment destination Fixed payTo per endpoint Dynamic payTo routing per request, including callbacks
Finding paid endpoints Out of band, read the docs Discovery extension, facilitators index endpoints and prices
Paywall Bundled Extracted into the @x402/paywall package
Backward compatibility n/a SDKs detect and serve v1 clients

Frequently asked

Questions people ask about x402 Protocol

What is x402?

x402 is an open protocol that lets a client pay for a web resource inside an ordinary HTTP request. A server answers an unpaid request with the 402 Payment Required status code and a JSON body listing its price, asset, network and address. The client signs a payment, retries with it attached, and gets the resource. It was built at Coinbase and is now governed by the Linux Foundation.

How does x402 work?

In four beats: the client requests a resource, the server replies 402 with machine-readable payment requirements, the client signs a payment and resends the same request with the signature in a header, and a facilitator verifies and settles the transfer on chain before the server returns the resource. No account, no API key and no checkout page is involved at any point.

What does x402 mean?

The name is a direct reference to HTTP status code 402, Payment Required, which was reserved in the original HTTP specification and left unused for about thirty years. The x prefix follows the convention of an experimental or extension standard built on top of an existing one. So x402 reads as the payment extension for HTTP 402.

Is x402 a crypto token?

No. x402 is a protocol specification, not a coin, and the x402 Foundation has issued no token. Tokens using the x402 name trade on exchanges and are not affiliated with the protocol or its governing foundation. If you are evaluating x402 as infrastructure for paid APIs or agent payments, nothing about it requires you to hold anything other than the stablecoin you settle in.

What is HTTP 402 Payment Required?

HTTP 402 is a client error status code that the original HTTP specification reserved for payment and never defined further, so for decades servers used 401 or 403 or a redirect instead. It is a real, allocated status code, just an unused one. x402 is the first widely adopted attempt to give it a standard, machine-readable meaning.

What is an x402 facilitator?

A facilitator is a service that sits between an HTTP server and a blockchain and does two things: verify that a signed payment payload is valid and settle the payment on the seller's behalf. It exists so a server can charge per request without its team managing wallets, gas or chain reorganizations. Coinbase runs one, and v2 lets others plug in.

What is the difference between x402 and MPP?

Both use HTTP 402 to signal that a request needs paying. x402 is blockchain-only and settles in stablecoins. The Machine Payments Protocol, co-authored by Stripe and Tempo, supports stablecoins, cards and Lightning, and includes idempotency, expiration and request-body binding in its base specification. x402 stays more permissionless; MPP reaches buyers who cannot hold crypto.

Does x402 support spending limits?

Not for the buyer. The maxAmountRequired field caps what a seller may charge for one call, which protects the seller, not you. The protocol holds no state about how much an agent has already spent, so a runaway retry loop is bounded only by the wallet balance. Per-agent budgets, counterparty caps and approval thresholds have to be enforced in a layer above the protocol.

What blockchains does x402 support?

x402 is chain-agnostic. The official description covers all EVM-compatible chains plus Solana and others, with USDC as the common settlement asset, and v2 identifies chains with CAIP identifiers so new networks need no custom code. Which chains you can actually use in practice depends on the facilitator you settle through, so check its supported list before committing.

How do I implement x402?

On the seller side you add payment middleware to your existing server that returns a 402 with your price and calls a facilitator to verify and settle, which is a handful of lines in the official SDKs. On the buyer side your client wraps its HTTP calls so a 402 triggers a signature and retry. Build against v2 headers and keep a v1 fallback.

Who owns x402?

Nobody owns it now. Coinbase created the protocol and contributed it to the Linux Foundation, which launched the x402 Foundation operationally on July 14, 2026 with 40 member organizations. Premier members include Visa, Mastercard, American Express, Google, AWS, Stripe, Adyen, Cloudflare, Shopify and Circle, so the specification is set by a governing body rather than a single vendor.

Keep reading

More explainers

ServiceNow AI Control Tower

ServiceNow AI Control Tower

ServiceNow AI Control Tower is the most complete agent inventory and risk console a large US enterprise can buy, and it now reaches across AWS, Google Cloud and Azure. We read the schema ServiceNow ships to developers to answer the one question the rollout meeting always ends on: can it stop an agent from spending money? It cannot, and the reason is written into the data model.

Read

Gemini Enterprise

Gemini Enterprise

Google did something in August 2026 that the other agent platforms have not done: it shipped a hard monthly spend cap that genuinely stops usage instead of emailing you about it. That deserves credit, and it also moves the interesting question one step along. A cap that stops something is only as useful as the thing it is scoped to, so we went and measured what Google can actually point that cap at, in the API model Google publishes for anyone to read.

Read

Salesforce Agentforce

Salesforce Agentforce

Agentforce is the largest agent platform any US enterprise is likely to already own, and it moved to consumption billing, which means the meter now runs on what your agents do rather than on how many seats you bought. That raises a finance question the rollout deck rarely answers: when an Agentforce agent is loose in production, what actually stops it spending. We went and measured the answer in Salesforce own published object model rather than guessing at it.

Read

AWS AgentCore

AWS AgentCore

Amazon shipped the missing piece in August 2026. Bedrock AgentCore Payments went generally available, and it is a real payments product: an agent can now hold a wallet, meet an HTTP 402, pay, and carry on reasoning without a human in the loop. So the question a platform lead has to answer stopped being whether AWS gives agents money and became a narrower, more awkward one: how much of a spend policy did AWS actually ship? We went and measured it, property by property, in the API model AWS publishes.

Read

Microsoft Agent 365

Microsoft Agent 365

Microsoft shipped a control plane for AI agents, and it is a good one. It gives every agent an identity, a registry entry, an owner, a sponsor and a Conditional Access policy. Then somebody in finance asks the obvious follow-up question: fine, but what stops the agent from spending money? This page answers what Agent 365 costs, what it governs, and what we measured when we went looking for a dollar amount anywhere in Microsoft's agent governance surface.

Read

QuickBooks MCP Server

QuickBooks MCP server

Connecting an accounting system to an AI assistant is now a ten minute job. Deciding what that assistant is allowed to do once it is connected is the part nobody writes about, and it is the part your controller will ask about first. This page compares what the official QuickBooks, NetSuite and Xero MCP servers actually hand a model, measured rather than summarized from marketing pages.

Read

Payment MCP Servers

payment MCP servers

Every large payment company shipped an MCP server in the last eighteen months, and almost every write-up of them is a setup tutorial. The setup is the easy part. The question worth answering before you connect one to a production account is narrower and much less comfortable: what, exactly, can the model on the other end of that connection do to your money?

Read

PayPal Agentic Commerce

PayPal Agentic Commerce

PayPal made a bet that most merchants would rather not implement a commerce protocol at all. Where Stripe and OpenAI shipped a spec for you to build against, PayPal shipped two products that sit on top of the checkout you already have, and then bought a company to make the catalog half work. That choice is the whole story: it explains why Agent Ready needs almost no engineering from you, why there is nothing for an agent to discover about your store on the open web, and why the thing PayPal will not do for you is the thing that gets expensive later.

Read

Shopify Agentic Commerce

Shopify Agentic Commerce

Shopify switched agentic commerce on by default, so your store is probably already selling to AI assistants whether or not anyone on your team configured it. Instead of restating the announcement, we checked something you can check too: on September 2, 2026 we requested the machine-readable capability file that Shopify publishes for real storefronts, on fourteen well-known US brand domains, and read what it exposes to an agent. Eleven answered correctly. The three that did not share one trait, and it is quietly costing them agent traffic.

Read

Web Bot Auth

Web Bot Auth

Web Bot Auth is the reason your agent either gets served or gets throttled with the scrapers. Almost everything written about it repeats the same architecture diagram, so we did something different: on September 1, 2026 we fetched the published key directories of more than twenty major AI operators and infrastructure vendors to see who is genuinely signing their traffic. Four were. The results are in the first table.

Read

Tempo Blockchain

the Tempo blockchain

Tempo is the payments chain Stripe and Paradigm built, and it shipped with a protocol that lets software pay for things on its own. It settles machine payments in under a second. It has nothing at all to say about whether your agent should have paid.

Read

AI Agent Governance

AI agent governance

Every agentic AI governance framework published so far governs the same four things: identity, tools, data and prompts. Not one of them carries a budget. Here is what the real frameworks say, which guardrails actually bind at runtime, and what to do about the last mile none of them reach.

Read

A2A Protocol

A2A Protocol

Most explanations of the A2A protocol stop at the sentence that agents can now talk to each other. That was true in April 2025 and it is no longer the interesting part. A2A shipped version 1.0 in April 2026 under Linux Foundation governance, it runs in production inside Azure AI Foundry and Amazon Bedrock AgentCore, and the questions engineers actually get stuck on are narrower: what an Agent Card commits you to, when to reach for MCP instead, and what happens the first time one of your agents has to pay another one for the work. That last question has a specific answer, and it is not in the core spec.

Read

Mastercard Agent Pay

Mastercard Agent Pay

Nearly every article about Mastercard Agent Pay is a retelling of one press release from April 2025, the one where Mastercard said AI agents would be able to shop with Agentic Tokens and named Microsoft as the first platform. That was sixteen months ago, and four more things have shipped since. Reading only the launch coverage leaves you with roughly a quarter of the picture, and the missing three quarters are the parts that decide whether you can actually put this into production.

Read

Visa Intelligent Commerce

Visa Intelligent Commerce

Almost everything written about Visa Intelligent Commerce is a retelling of the April 2025 announcement, when Visa said AI agents would be able to pay with a Visa credential. Three more things have shipped since, including an open agent-identity protocol built with Cloudflare that most coverage does not mention at all. This page is the current version, checked against Visa’s own developer documentation and newsroom in August 2026.

Read

Stripe agentic commerce

Stripe agentic commerce

Most writing about Stripe and agentic commerce is still a retelling of the September 2025 launch week, when Stripe and OpenAI shipped Instant Checkout and published the Agentic Commerce Protocol together. Stripe has built a good deal more since then, and some of it points in a direction the launch coverage never anticipated. This page is the current version, checked against Stripe’s own documentation in August 2026.

Read

ChatGPT Instant Checkout

ChatGPT Instant Checkout

Almost every guide to ChatGPT Instant Checkout still reads like it was written the week it launched, walking merchants through how to apply and what the fee will be. OpenAI changed course in March 2026. Here is the accurate version: what Instant Checkout was, what the numbers actually looked like, what replaced it, and which parts of the stack are still very much alive.

Read

Google AP2

Google AP2

Most guides to Google AP2 still describe an Intent Mandate and a Cart Mandate, because most of them are rewrites of the September 2025 launch post. The specification moved. Here is what the Agent Payments Protocol actually defines today, and the one question it deliberately does not answer.

Read

Human in the loop AI

Human in the Loop AI

Every guide to human in the loop AI describes the same shape: the agent pauses, a person decides, the agent continues. The shape is right. What almost none of them ask is a harder question, which is where the pause is enforced, because a pause written into the agent's own code is a pause the agent is trusted to honor.

Read

AI agent cost

AI Agent Cost

Every cost guide for AI agents answers the same two questions: what does it cost to build, and what does it cost to run. Both are answerable, and both are on somebody's invoice. The third question is the one that ends up in a variance report, because the agent also spends your money, and nobody sends you a bill for that.

Read

Agentic checkout

Agentic Checkout

Nearly every guide to agentic checkout is written for the merchant who wants to receive these orders. Far fewer are written for the company whose agents are placing them, which is odd, because agentic checkout quietly removes the one screen where spending used to get a second look.

Read

API monetization

API Monetization

Most guides to API monetization argue about which pricing model wins. The harder question in 2026 is who is calling. An API priced for a signed-up developer with a key behaves very differently when the caller is an agent that showed up once, wants one record, and has no account.

Read

AI procurement agents

AI Procurement Agents

Every major procurement suite shipped agents during 2026. Almost none of them answer the question your controller will ask first, which is what happens when the agent is wrong about a purchase and the money has already moved.

Read

Agentic payments

Agentic Payments

Agentic payments move money with no human at the checkout. The rails to do it all shipped during 2026. The part most teams have not solved is deciding, before the money moves, whether the agent was allowed to spend it.

Read

AI agent monetization

AI Agent Monetization

Every AI agent company is rewriting its price list. The models that survive are metered. The ones that quietly fail are the ones where nobody measured what a single task costs to serve.

Read

Agent payment platforms

AI Agent Payment Platforms

Five different kinds of product now call themselves an AI agent payment platform, and they solve five different problems. Picking the wrong category is the expensive mistake, not picking the wrong vendor inside a category.

Read

Universal Commerce Protocol

the Universal Commerce Protocol (UCP)

Google and Shopify shipped UCP as an open standard so an AI agent can check out at any merchant that supports it. Here is what the specification actually defines, where it is live for US buyers, and the one thing it deliberately leaves to you.

Read

MCP Payments

MCP Payments

MCP payments are how an AI agent discovers a payment tool and calls it to move money. The catch: the Model Context Protocol carries the tool call, not the spending decision, so nothing in the stack asks whether the purchase should have happened.

Read

Visa Intelligent Commerce vs Mastercard Agent Pay

Visa Intelligent Commerce vs Mastercard Agent Pay

Visa Intelligent Commerce and Mastercard Agent Pay are the two big card networks racing to let AI agents pay. They take different routes to the same idea, and neither one decides whether a given purchase should have happened.

Read

Agentic Commerce Protocol

the Agentic Commerce Protocol

ACP is the open standard behind agentic checkout in ChatGPT. It tells a merchant how to sell to an AI agent. It says nothing about whether your agent should have made the purchase.

Read

AP2 vs ACP vs x402

AP2 vs ACP vs x402

AP2, ACP, and x402 are the three standards shaping how AI agents pay. They solve different layers of the problem, and most real systems will touch more than one.

Read

Machine payments protocol

Machine payments protocol

As software starts paying software, machine payments protocols define how value moves without a human at the keyboard. The harder question is how to keep that spending governed.

Read

Know Your Agent (KYA)

Know Your Agent

KYA, or Know Your Agent, extends the idea of customer due diligence to autonomous software. When an agent spends, you need to know which agent, on whose authority, and under what limits.

Read

Keep agent spending governed

Add policy, hard limits, human approval, and an immutable audit trail across any protocol or rail. Start in the sandbox today.

Never moves money without policy