Amazon Blocks Muse from Shopping on Amazon.com
Amazon has cut off Meta’s AI agent Muse from shopping on Amazon.com. It says Meta did not arrange access in advance and that the agent does not transparently identify itself while browsing the store.
Amazon is blocking Muse from shopping on Amazon.com. On the evening of September 20 Pacific time, users of Meta’s AI agent began seeing a message when attempting to make a purchase stating that access by an unauthorized AI agent violates Amazon’s terms of use. This is a significant change shortly after the service launched in the US: according to available reports, shopping on Amazon.com is no longer available to Muse.
Amazon confirmed to GeekWire that it cut Muse off from shopping in its store. It also said that it had asked Meta to remove Amazon from this feature. According to available reports, Meta did not immediately issue a public statement after the block.
Why Amazon is blocking Muse
Amazon says Meta did not inform it in advance about Muse’s access. According to the company, the agent does not identify itself while browsing the store, and its activity could pose risks to user privacy and security.
The dispute also centers on how more autonomous shopping tools should access online stores. Although a user may authorize an agent to search for or complete a purchase, the store operator may require such a tool to identify itself transparently and obtain authorized access.
In an earlier letter to Perplexity, Amazon said that AI agents must be identifiable when accessing its store and that Amazon wants to decide whether to allow them such access. The current action against Muse therefore fits with this publicly communicated position.
Meta says credentials are stored securely
When introducing Muse, Meta stated that the agent does not see users’ passwords or payment details. The credentials are supposed to be stored in a secure environment referred to as Muse Secure VM. The company also said that Muse is supposed to require user approval for sensitive actions.
Amazon’s technical claim that Muse captures and stores customer login credentials has not been independently verified. It also conflicts with Meta’s description, according to which the agent does not have access to passwords. Based on the available information, it is therefore impossible to determine exactly how Muse handled login credentials when attempting to make a purchase on Amazon.com or which specific security mechanisms Amazon considered problematic.
Meta presented Muse as a personal AI agent designed for multistep tasks, such as working with email, calendars, reservations and purchases. The block therefore limits one of the service’s practical functions, at least with regard to Amazon.com.
The scope of the block remains unclear
It has not been confirmed whether the measure affected all user accounts, all Muse interfaces or every step of the shopping process. It is also unknown whether Meta will modify the agent so that it meets Amazon’s requirements, or whether the companies will reach a technical or commercial agreement on authorized access.
It also remains an open question whether Amazon will publish more detailed technical reasons for the block or take legal action. Meta’s public response and any changes to how Muse works with Amazon.com will be particularly important.
The case illustrates a specific conflict between shopping AI agents and the platforms on which they are supposed to perform tasks. In this case, Amazon made clear that a user’s authorization alone is not considered sufficient grounds for an external agent to access its store without transparent identification and the operator’s consent.
Sources
- GeekWire – Reports the message shown to users and Amazon’s quoted statement about blocking Muse, its communication with Meta and the stated reasons.
- Meta Newsroom – Documents Meta’s claims about Muse Secure VM, credential storage, the agent’s lack of access to passwords and approval of sensitive actions.
- The Verge – Independently confirms that the blocking message appeared on September 20 and that Amazon provided GeekWire with a statement.
- Amazon – letter to Perplexity – Provides context on Amazon’s longstanding policy: it requires transparent identification of AI agents and the ability to decide whether to allow them access to the store.
Verified and updated: September 21, 2026 15:23
