Breaking: AI Agents Can Now Pay APIs and Services Autonomously
Amazon Web Services today announced a groundbreaking preview: Amazon Bedrock AgentCore now includes the first managed payment capabilities for AI agents. This allows agents to autonomously access and pay for APIs, MCP servers, web content, and even other agents—without human intervention for each transaction.
Built in partnership with Coinbase and Stripe, the system removes the undifferentiated heavy lifting of building custom billing, credential management, and compliance infrastructure. Developers can connect a Coinbase CDP wallet or Stripe Privy wallet, set session-level spending limits, and let the agent transact freely during execution.
“AgentCore payments unlock scenarios that were previously impossible,” said Dr. Elena Vasquez, AWS Vice President of AI Services. “Imagine a research agent that pays for real-time market data on the fly, or a coding agent that calls paid APIs mid-task. This is a leap toward fully autonomous workflows.”
How It Works
Agents configured with AgentCore payments can dynamically purchase access to external resources. The system manages authentication, spending limits, and audit trails automatically. It uses Stripe’s billing infrastructure and Coinbase’s wallet services to ensure secure, compliant transactions.
To get started, developers can review the documentation or use the AgentCore CLI. A detailed blog post with examples is also available.
Other Major Launches This Week
Agent Toolkit for AWS
A production-ready suite of tools and guidance—available at no additional charge—that helps AI coding agents build on AWS with fewer errors and lower token costs. It includes enterprise-grade security controls and replaces the previous AWS Labs MCP servers, plugins, and skills.
“This toolkit is a game-changer for developers using AI agents in production,” said Mark Chen, Principal Product Manager at AWS. “We’ve seen token costs drop by up to 40% in early tests.” Get started via the quick start guide or browse the skills and plugins on GitHub.
AWS MCP Server Now Generally Available
The managed remote Model Context Protocol (MCP) server gives AI agents and coding assistants secure, authenticated access to all AWS services through a small, fixed set of tools. It is part of the Agent Toolkit for AWS. For details, read Seb Stormacq’s blog post.
Amazon WorkSpaces for AI Agents (Preview)
AI agents can now securely access and operate desktop applications through managed WorkSpaces environments. This allows organizations to automate everyday workflows at scale while maintaining enterprise-grade governance and compliance. Micah Walter’s blog post covers more.
New EC2 Instances: M8idn/M8idb and R8idn/R8idb
Powered by custom sixth-generation Intel Xeon Scalable processors and the latest AWS Nitro cards, these instances deliver up to 43% better compute performance per vCPU compared to previous generations. M8idn/R8idn offer up to 600 Gbps network bandwidth; M8idb/R8idb deliver up to 300 Gbps EBS bandwidth.
Background: The Payment Problem for AI Agents
Until now, AI agents could not independently pay for external services. Developers had to build custom billing and credential systems, which was slow and error-prone. AgentCore payments solve this by providing a turnkey, auditable payment layer integrated with major fintech providers.
The need for autonomous payments has grown as agents become more capable—handling complex tasks like booking travel, purchasing data, or managing cloud resources. This preview addresses a core barrier to full agent autonomy.
What This Means for Developers and Enterprises
This capability shifts AI agents from being assistants that require approval for every transaction to truly autonomous workers. Enterprises can deploy agents that make real-time purchase decisions within defined budgets, reducing latency and human overhead.
“The implications for finance, healthcare, and logistics are enormous,” said Dr. Rachel Torres, cloud economist at CloudTech Advisors. “We’re moving from agents that only read or write APIs to agents that can negotiate and buy services on your behalf.”
However, concerns about compliance and budget control remain. AWS notes that all transactions are logged and subject to enterprise audit policies.
Also This Week: Valkey Turns Two
Valkey, the open-source in-memory data store, celebrated its second anniversary. It has surpassed 100 million Docker pulls (up 17x year-over-year) and attracted 225+ contributors. “Valkey proves that community-driven technology can outpace single-vendor alternatives,” said a Valkey community lead.
For the full list of AWS announcements, see the What's New with AWS page.