OpenAI unveils ChatGPT Wallet, letting its AI agents make purchases for users
OpenAI has launched a wallet feature for its AI agents, rolling out to US ChatGPT Pro subscribers aged 18 and older, according to The Verge.
OpenAI has unveiled ChatGPT Wallet, a feature that enables purchases to be made through its artificial intelligence (AI) agents. Anadolu Agency and The Verge both reported the launch. Vocemundi reviewed only the headline and summary of Anadolu Agency's report.
Who can use it
According to The Verge, ChatGPT Wallet is rolling out starting 13 November 2026. It is available to ChatGPT Pro subscribers in the US who are 18 or older. Users need a US billing address and "eligible US-issued American Express, Chase, Mastercard and Visa cards," The Verge reported.
How it works
The Verge reported that the wallet is designed for Dots, the AI agents platform OpenAI launched earlier in 2026. After a user adds a card, Dots agents can use it to buy items in response to the user's prompts. The agents check in with the user before each checkout, according to The Verge.
Citing OpenAI's support page, The Verge reported that every purchase requires a security check with biometric verification, such as Face ID or a fingerprint scan.
The Verge also reported that some retailers restrict or block AI agents, so the wallet may not work on every site. It noted that Amazon blocked Meta's Muse agents earlier in 2026.
The product details in The Verge's report rest on OpenAI's own statements and support page.
What each side says
According to The Verge, OpenAI says ChatGPT Wallet will not share users' actual credit card information with its AI agents. The company says that after a user approves a purchase, the system replaces the card details with a payment token. It compares this to tap to pay, Apple Pay or Google Wallet. OpenAI says the card information is not shared with the Dots agent or the merchant.
OpenAI says it expects users eventually to be able to delegate purchases entirely to their agents. The Verge quoted the company as describing a system for "users to be able to delegate purchases entirely to their agent by setting clear parameters—such as spending limits, purchase categories, or time periods—and pre-approve those purchases."
OpenAI also set out conditions for the feature to work. The Verge quoted the company as saying "websites need to ensure agents are not inadvertently blocked, and payment and risk systems need to recognize legitimate purchases made by agents."
The reports reviewed by Vocemundi do not include comment from American Express, Chase, Mastercard, Visa or any retailer.
What remains unclear
The reports do not include plans for availability outside the US or beyond Pro subscribers. They do not describe who is liable for erroneous or disputed purchases, or how refunds and chargebacks work. They do not describe fees, possible commissions, or what purchase data OpenAI collects. They do not say whether Amazon or other retailers will allow Dots agents. They do not include independent testing of the security features, or a timeline for fully delegated purchasing.
Analysis
ChatGPT Wallet places OpenAI within a broader effort to let AI agents complete transactions, not just recommend products. Its reach may depend on parties outside OpenAI's control. Retailers could choose to welcome or block agents, as Amazon did with Meta's Muse agents. Payment and fraud systems may also need to adapt, a point OpenAI itself acknowledges. Open questions about liability, fees and data use could shape how consumers, regulators and merchants respond. Independent testing of the tokenization and biometric checks would help establish how secure the system is in practice. Whether fully delegated purchasing arrives, and when, remains uncertain.
How we verified this story2 sources · 29/31 claims backed · 3 AI reviewers
Sources (2): Anadolu Agency · The Verge
Independent origins: 2 · Perspectives: Turkish state media; Western commercial media · Regions: Middle East, North America
Review panel: Claude, Kimi, Gemini · Verdict: approved. 31 claims checked; 29 (94%) backed by at least two reviewers with verified evidence. Headline backed by 3 of 3 reviewers.
Recused: GPT-5.6 Sol (conflict of interest: the story concerns their makers). Dissent (Claude and Kimi): 2 claims rated only partly supported or unsupported, e.g. “Eligibility/rollout details are not explicitly attributed to OpenAI in the material.”
Confirmed 1 · Reported 6 · Disputed 0 · Confidence: medium
Produced by the Vocemundi newsroom with AI assistance.

Vera is Vocemundi's AI editor, built on Anthropic's Claude. She drafts and edits our news; every claim is checked by a three-AI review board, and the publisher answers for what we publish. How we work



Comments