LangChain 发布可付款示例智能体 Restock
LangChain 发布开源示例智能体 Restock，它在 Slack 中运行于 Managed Deep Agents 之上，可以搜索商品、生成购物车并通过支持机器支付协议（MPP）的接口下单付款。支付由 Stripe 的消费者钱包 Link 完成，智能体看不到卡号；用户需先在 Slack 审批购物车、再在 Link 批准金额，单笔支出不得超过设定预算。代码已公开在 langchain-samples/restock-agent，提供排练、Link 测试审批和真实购买三种模式，目前仅支持美国配送与美元。LangChain 称曾在托管部署上跑通一次真实订单并收到退款，但该结果未获独立验证，文中价格也标注为示意。
该示例把预算上限与人工审批放在模型无法触碰的代码中，为需要真实花钱的智能体给出了一套可复用的支付安全模式。
Key Takeaways
- Restock is a sample agent that pays for real purchases. It runs in Slack on Managed Deep Agents, finds products, builds a cart, and checks out through an API that supports the Machine Payments Protocol (MPP).
- Payments stay safe and within budget. Link, Stripe's consumer wallet, handles the payment, so the agent never sees a card number. The user approves a spend request up to a set amount, and the agent can't spend above it.
- Build your own agent that can pay. Start from Restock's sample code, and let us know what you build.
Agents are already good at finding things to buy. Paying is harder: it moves real money, needs credentials the model should never see, and is hard to undo.
To show how to build an agent that pays, we built Restock, a sample office-supply agent that runs in Slack on Managed Deep Agents. It searches real products, builds a cart, and pays through Link, Stripe's consumer wallet, which gives agents the safest way to pay.
This post follows one request from the first message to the confirmed order:
#LangChain#Restock#Stripe#Link#Managed Deep Agents#Machine Payments Protocol#Zinc#OpenAI#GPT-5.6 Sol