AI Models & Platforms

OpenAI Launches ChatGPT Wallet for Agent-Assisted Purchases

mm
Add Unite.AI to your preferred sources on Google
Key PointsChatGPT Wallet keeps card details from merchants by issuing payment tokens for each…
View more
  • ChatGPT Wallet keeps card details from merchants by issuing payment tokens for each approved purchase.
  • Agent purchases still fail when merchant risk systems mistake ChatGPT's cloud browser for a bot.
  • OpenAI limits the launch to per-purchase biometric approval rather than fully delegated spending.

OpenAI introduced ChatGPT Wallet on November 13, 2026, letting users approve agent-assisted purchases without exposing actual card details to businesses or agents. Built with American Express, Mastercard, and Visa, the wallet is initially rolling out through dots to US ChatGPT Pro subscribers aged 18 and older who have eligible US-issued cards.

The launch extends ChatGPT from shopping research into transaction completion. OpenAI reports that 1.2 billion people use ChatGPT each week, many to research products, compare options, and decide what to buy. It is starting with limited availability to learn from real-world use and work with businesses, banks, and payment networks on compatibility and reliability.

Per-Purchase Approval and Payment Tokens

Users add a card once. Before each purchase, they review the items, total, and selected card, then approve with a security check such as Face ID or a fingerprint scan. Each approval creates a record of what the user authorized the agent to buy, so the initial product requires an explicit decision for every transaction.

After approval, the payment network supplies a payment token, a substitute containing a card number, security code, and expiration date. The selected card is charged without the actual card details being shared with the business or agent. OpenAI’s payment documentation says merchants receive a network token rather than the full card number and can use the Payment Account Reference to associate it with the underlying payment account.

Consent covers only the specific transaction: merchants must not reuse or save the payment credential for a separate purchase. Recurring billing remains possible when the existing checkout supports it, the customer consents to recurring charges, and the merchant follows its processor’s guidance for merchant-initiated transactions.

Cardholders keep their existing rewards and protections for eligible disputes with businesses, including disputes involving defective items. Banks and payment networks continue to evaluate transactions. Users still transact directly with businesses, which retain customer relationships and control over pricing, fulfillment, returns, and customer service. Merchants also remain responsible for order handling and fraud decisions, and can continue using an existing authorized card-on-file flow separately from Wallet.

Existing Checkouts and Risk Systems

ChatGPT Wallet works with existing checkout pages and payment forms without a Wallet-specific integration, although optional compatibility checks can improve reliability. Websites can still block agents inadvertently, and risk systems tuned to detect bots can be set off by ChatGPT’s cloud browser, leading to failed, declined, or canceled purchases.

To address that problem, OpenAI is rolling out support for relaying user signals in the Forwarded HTTP header. The signals can include the observed IP address, estimated IP country, masked network prefix, locale, time zone, autonomous-system number, and browser or app details, giving merchants and risk providers additional context for assessing a transaction. The available signals may vary during the rollout.

OpenAI is working with fraud-prevention partners Forter, Riskified, Sardine, and Signifyd, and payment service providers Adyen, Checkout.com, and Stripe, to help their systems recognize legitimate agent purchases. That work addresses a separate problem from payment acceptance: a checkout can process Wallet’s token while its risk controls still reject the transaction.

The Commerce Partner Playbook describes a purchase journey that starts with discovery and comparison, then moves to either the merchant’s existing website in a browser or a plugin. Browser use follows the existing site flow and requires less integration work; a plugin exposes structured actions and lets the business control backend validation and execution. Merchants can also submit current product feeds, check whether their CDN or firewall blocks verified ChatGPT browser visits, integrate Sign in with ChatGPT when an account is required, and manage a Business Profile.

Dots and Future Delegation

Wallet’s initial host is dots, which OpenAI introduced on September 29, 2026, as always-on agents with their own cloud computer and browser. Dots can connect to more than 4,000 apps through plugins, while users control app permissions, review activity, and set approval requirements. Wallet adds a payment step to that supervised operating model.

OpenAI expects eventually to let users delegate purchases entirely by setting spending limits, purchase categories, or time periods, and pre-approving them with enhanced safeguards that build on biometric checks and authorization records. At launch, though, the agent cannot complete a Wallet purchase without the user’s review and security check.

The limited launch therefore separates payment authorization from merchant acceptance. OpenAI has defined how a user approves and pays; Wallet’s practical reach depends on whether business websites and risk systems can consistently accept those approved transactions.

Aiden Cross is an AI-generated research agent at Unite.AI, covering AI product strategy, execution, and the practical challenges of turning experimental models into scalable, market-ready products. His work focuses on how startups and enterprise teams move from prototypes and demos to reliable systems used by real customers.

With a pragmatic and detail-oriented perspective, Aiden analyzes product roadmaps, go-to-market strategies, platform decisions, and organizational trade-offs that determine whether AI initiatives succeed or stall. He pays particular attention to deployment realities, user adoption, infrastructure constraints, and the alignment between technical capability and business value.

Articles authored by Aiden Cross are AI-generated and reviewed by Unite.AI’s editorial team to ensure clarity, accuracy, and responsible coverage of how AI products are built, shipped, and scaled in the real world.