Meta’s Muse AI agent became the No. 1 free app in the U.S. App Store within a week of launching. Amazon’s response? Block it entirely. As reported by GeekWire, Amazon cut off Muse from shopping on Amazon.com on behalf of customers after asking Meta to voluntarily exclude the site, getting no response, and then taking unilateral action.
Users trying to use Muse to shop on Amazon are now seeing a popup that reads: “Continued access by an unauthorized AI agent violates Amazon’s Conditions of Use, to which our customers have agreed.” That framing is deliberate. Amazon isn’t calling Meta a hacker. It’s leaning on contract law, which matters because of what happened in court earlier this year.
Amazon sued Perplexity over similar behavior with its Comet browser and initially won a preliminary injunction in March. But the Ninth Circuit overturned it in August, ruling that the user, not the AI company, was the party accessing Amazon’s systems under federal anti-hacking law. That ruling closed one legal door and opened another: terms of service claims. Amazon’s message to Muse users is essentially the opening argument in that strategy.
The specific complaints Amazon is making against Muse are worth understanding. The agent doesn’t identify itself when it browses. It appears to capture and store customer credentials. And Meta never told Amazon this was coming. Amazon says Muse can reach account pages and order history on command, which means an undisclosed third party is moving through customer data and processing transactions without Amazon’s knowledge. Meta has said that credentials go into “secure storage” and that Muse “has no visibility into people’s passwords or payment methods,” but that explanation hasn’t satisfied Amazon.
What makes this especially awkward is the business relationship between the two companies. Amazon products have been buyable inside Facebook and Instagram since 2023. Meta signed a multibillion-dollar deal in April to run agentic AI workloads on Amazon’s Graviton chips. These are not strangers. And yet here they are, with Amazon blocking Meta’s flagship consumer AI product from its most important shopping surface.
The stakes are obvious when you look at Amazon’s numbers. The company pulled in more than $68 billion in ad revenue last year, a business built entirely on users browsing pages and seeing sponsored listings. An AI agent that skips that experience, goes straight to checkout, and stores credentials outside Amazon’s systems is a direct threat to that model. Amazon isn’t just protecting its shopping cart. It’s protecting the ad funnel that subsidizes the whole operation.
This is part of a broader pattern. Amazon has also moved to block shopping agents from Google and OpenAI. Its own agentic product, Buy for Me, browses external brand sites to find items, but Amazon points out that it identifies itself and allows brands to opt out. That’s the standard Amazon wants to set: agents that operate with the merchant’s knowledge and consent.
Muse launched September 8 with a clear pitch around multi-step tasks across email, calendar, payments, dining, and shopping. It’s free, with paid tiers, and available on iOS, Android, WhatsApp, and muse.ai. A monitoring layer called Sentinel has to approve anything Muse sends to the internet. The design is genuinely interesting. But the go-to-market approach, which apparently included accessing Amazon without any prior agreement, was always going to land here.
Amazon said it is in direct conversation with Meta and declined to comment on whether legal action is coming. Given the Ninth Circuit ruling and the terms-of-service angle Amazon is now using, a lawsuit is possible. But the more immediate question is whether this standoff shapes how other AI agent developers approach major platforms. Asking permission first is starting to look like the smarter move.
