
Anchor Browser Integrates AgentCore Payments for Autonomous x402 Transactions
Anchor Browser now integrates with Amazon Bedrock AgentCore payments, enabling AI agents to autonomously pay for x402-protected pages and APIs. Wallet keys stay out of the runtime, and budgets are enforced at the infrastructure level.
Agents stall at paywalled resources
AI agents increasingly need access to web resources priced per request, such as premium datasets, research APIs, and paywalled publishers. When an agent hits one of these resources, the server returns HTTP 402 Payment Required. Without a payment mechanism, session budget, or programmatic way to satisfy the request, the task stalls and requires manual intervention.
Anchor Browser with Agentic Payments handles machine-to-machine payment workflows on behalf of the agent. Developed with the AWS Generative AI Innovation Center Partner Agent Factory, it is available on AWS Marketplace. It pairs Anchor Browser, a cloud-based browser automation solution for AI agents, with Amazon Bedrock AgentCore payments, a managed capability for agents to pay for APIs, MCP servers, and web content autonomously.
How the payment flow works
The x402 protocol is an open standard for HTTP-based machine-to-machine payments. A server responds with 402 Payment Required along with the amount, accepted payment methods, and payment address. When an agent in Anchor Browser hits such a resource, AgentCore payments authorizes and completes the transaction within the configured budget, and the workflow continues without interruption.
Anchor Browser creates a dedicated AgentCore PaymentSession for each session started with x402 enabled. The browser reads the 402 challenge, matches the payment method against the linked instrument, and AgentCore verifies the budget and generates the cryptographic proof. No key material enters the browser or agent runtime. The browser attaches the signed proof as PAYMENT-SIGNATURE for x402 v2 or X-PAYMENT for x402 v1 and retries the request exactly once. The merchant verifies, settles, and returns the protected resource.
At the center is the Payment Manager, the top-level resource coordinating payment operations for the AWS account. A PaymentConnector integrates it with a wallet provider, either Coinbase Developer Platform (Coinbase CDP) or Stripe Privy. A PaymentCredentialProvider stores provider credentials in AgentCore Identity, and a PaymentInstrument identifies the wallet address funds are drawn from. For endpoint discovery, AgentCore offers a curated Coinbase x402 Bazaar MCP server through AgentCore Gateway.
Enabling payments requires setting session.x402 to true when creating a session; it defaults to false. Once enabled, the browser handles 402 challenges, signing, and retries transparently for navigation, fetch, and XHR requests, with no wallet key or payment header in automation code. Existing Playwright, Puppeteer, and Chrome DevTools Protocol workflows continue to function unchanged. For paid API endpoints, a built-in x402 tool handles POST, PUT, and other HTTP methods the same way within the browser context.
Security, budgets, and billing
The integration moves the payment stack into managed infrastructure. The Anchor Agent holds only a budget session and the x402 tool, while authentication, wallet secrets, spending limits, observability, and proof signing reside in the AgentCore payments layer. Signing and settlement happen inside AgentCore, so the runtime never has access to key material that can move funds.
Budgets are enforced at the infrastructure level before any payment is authorized, and payment attempts, approvals, denials, and settlements are visible through Amazon CloudWatch. When a session hits its spending limit, AgentCore denies the payment. Anchor Browser also settles third-party x402 charges in credits on the same invoice as browser session time, with no separate crypto custody or on-chain settlement to reconcile.
Sources: AWS Machine Learning Blog
SiTech — AI-powered web development
We build fast, modern websites and bring AI into real business workflows. Have a project or a question? We'd love to help.