What is Muse, and why did Amazon block it?
Muse is Meta's AI agent, built to handle multi-step tasks like email, calendar bookings, dinner reservations and shopping, all from one app on iOS, Android, WhatsApp or the web.
Meta launched Muse on September 8, 2026. Amazon started blocking it twelve days later, on the night of September 20. Shoppers trying to use Muse on Amazon now see a pop-up: "Continued access by an unauthorized AI agent violates Amazon's Conditions of Use, to which our customers have agreed."
Amazon gave three reasons. Meta never asked permission for Muse to access Amazon's site. Muse doesn't identify itself as an AI agent while it browses, so Amazon can't tell it apart from a person clicking around. And Amazon says Muse appears to capture and store customer login credentials, which it treats as a security risk. Amazon says it acted only after Meta refused a direct request to pull the agent back.
Is Meta's security defense actually true?
Meta says Muse never sees a user's password or payment details in the first place, so the credential-capture concern is overstated.
According to Meta, Muse runs on a dedicated cloud computer rather than the user's own device, and it asks for approval before completing a purchase. When a user connects an account, the login details go into secure storage that Muse can use to act on the site without ever displaying or reading the raw credentials itself.
Both explanations can be true at once. Muse can be built with real safeguards, and Amazon can still have a legitimate objection to any automated tool logging into customer accounts at scale without a formal agreement first. Security researchers who don't work for either company haven't independently verified either version, so treat both as claims for now, not settled facts.
Why does a $68 billion number matter more than the security dispute?
Because Amazon's advertising business depends on shoppers scrolling its own pages, and an AI agent that shops for them skips that step entirely.
Amazon pulled in roughly $68 billion in advertising revenue last year. That money comes from brands paying to place sponsored listings in front of shoppers who are actively browsing and comparing products. If a Muse-style agent just grabs the item a user asked for and checks out, none of those sponsored listings get seen. The shopper never scrolls past them. Amazon isn't only defending account security here. It's defending the attention its whole ad business is built on.
That's also why the "unauthorized agent" language matters beyond this one product. Amazon isn't saying AI shopping agents are banned forever. It's saying it decides which ones get in, and on what terms, the same way a store can choose which delivery services are allowed to walk through its front door.
| Agent | Company | Amazon's move |
|---|---|---|
| Comet browser | Perplexity | Sued in November 2025 |
| Shopping agent | Blocked from Amazon's site | |
| Shopping agent | OpenAI | Blocked from Amazon's site |
| Muse | Meta | Blocked, September 20, 2026 |
Nexairi Dispatch
Get the next AI move before it turns into a workflow problem.
Join the free newsletter for concise AI news, practical checklists, and the decisions practitioners need to make next.
Free. No spam. You will also get the Nexairi Dispatch.
How did the market react, and does that change Amazon's calculation?
Meta's stock jumped 11.4% on September 21, its best single day in roughly a year, with Muse's strong reception cited as a key driver behind the move.
Zoom out further and the move looks even bigger. Meta shares climbed roughly 25 percent over the trailing month, from about $546 in mid-August to about $682 by mid-September, a run that coverage largely credits to Muse's launch and rapid download growth. Within about a week, Muse became the top free app in Apple's U.S. App Store, ahead of ChatGPT.
None of that popularity changes Amazon's legal footing. But it does raise the stakes on both sides. Every day Muse stays blocked from Amazon is a day Meta can point to as proof that a major platform is limiting a product investors are already rewarding. And every day Muse keeps growing without Amazon's cooperation is a reminder to Amazon that the demand for agent-based shopping isn't going away just because one retailer says no.
Does a customer's approval matter if the platform says no?
Not under current terms of service. A shopper's approval doesn't override the rules a website sets for who, or what, gets to access its pages and checkout.
This is the part of the story that reaches past shopping. The same question applies to booking a flight, moving money between accounts, renegotiating a cable bill or scheduling a doctor's visit. If every company gets to decide, on its own terms, whether an outside AI agent may act on a customer's behalf, then "AI agents can do this for you" quietly becomes "AI agents can do this for you, except everywhere the platform decides they can't." That gap between what an agent is technically capable of and what it's allowed to do is where this fight is actually happening.
What this means if you're building or buying with AI agents
If your business is evaluating agentic AI tools, whether for procurement, travel booking or customer-facing shopping, this dispute is the preview of a recurring cost. An agent that works great in a demo can still get blocked the moment a platform decides it doesn't like the access pattern, and you may not find out until it stops working mid-workflow. Ask any vendor pitching an AI agent how it identifies itself to third-party sites, and whether it has any formal agreements with the platforms it depends on, not just a technical claim that it "can" access them.
For consumers, the useful takeaway is smaller but concrete. If Muse or a similar agent stops working on a site you use often, the reason is more likely a business dispute over data and access than a bug in the app. Check whether the platform, not the agent maker, put up the block before assuming something's broken on your end.
The question worth watching isn't who wins this specific standoff. It's whether Amazon, Meta and the rest of the industry settle into some kind of formal access agreement, the way apps eventually negotiated API terms with each other, or whether every major platform ends up building its own wall against every other company's agent. One path leads to AI agents that actually work across the sites you use. The other leads to a shopping experience that depends entirely on whose agent you picked.
Sources
- TechCrunch: Meta's AI Agent Has Been Blocked From Using Amazon.com
- GeekWire: Amazon Blocks Meta's Muse AI Assistant in New Standoff Over Agentic Shopping
- Bloomberg: Amazon Blocks Meta's Muse AI Agent From Its Retail Site
- Yahoo Finance: Meta Jumps 11% As Muse Shines and Investors Show an Appetite for Advancing AI
- TradingKey: Meta Shares Surge Over 8% Ahead of Connect 2026 as Muse Becomes New AI Variable
Related Articles on Nexairi
Jim Smart is the founder and editor in chief of Nexairi. A Business Intelligence Developer with experience building data systems for Verizon, U.S. Army operations, and enterprise finance teams, Jim spent years turning complex data into decisions that executives could act on — dashboards, forecasting models, and automation pipelines across telecom and government contracting. He founded Nexairi to apply that same clarity to AI: making emerging technology understandable and actionable for the operators, accountants, and business owners who need it most. Jim holds GenAI certifications from the University of South Florida Bellini College of AI and completed Springboard's Data Science Career Track.
