Twelve days after the launch, Amazon blocked Meta's agent at checkout and cited hidden browsing and unauthorized credential processing.
From Sunday, September 20th to the evening, any user who tries to conclude a purchase on Amazon via Meta's personal agent, Muse, encounters an error message: the continued access of an unauthorized agent violates the terms of use of the site. Meta had launched Muse on September 8, presented as an assistant able to send emails, book trips, fill out forms and pay; A week later, the app was at the top of the American App Store's free ranking.
Amazon argues three grievances: no prior agreement was made with Meta, the agent does not announce himself as such when he is browsing the site, and it seems to collect and then retain customer identifiers — potentially access to account pages and order history. Meta, who did not react immediately, had assured that Muse had no visibility on passwords or means of payment.
The episode is part of a constant line: Amazon has already closed the door to Google and OpenAI agents, and is prosecuting Perplexity. The substantive issue is no longer technical but commercial — which controls the customer relationship when it is software that runs the errands.
Read the full article on the original source
Go to the source →