News
News

Is your bot protection still letting Meta's Muse in? The AI agent just changed its ID

Short answer

Maybe not. Between 7 and 9 November 2026, Meta's Muse agent changed the ID it uses to prove who it is, from meta.com to muse.ai. If your bot protection only trusts the old one, Muse's visits may now look like unknown bots and get blocked. That matters because Muse now makes most of the AI agent requests that bot-protection companies see.

News cover: grey network cables twisted together on a dark background, under the headline asking whether bot protection still lets Meta's Muse agent in.
On this page

    Key facts

    • Muse signs some of its requests with Web Bot Auth, a cryptographic ID that copycat bots can't fake.
    • From around 7 November 2026 Muse started signing as muse.ai; by 9 November it had stopped signing as meta.com, according to DataDome's threat research team.
    • As of 10 November, about half of Muse's requests still arrived without a signature, DataDome says.
    • In October, Muse made 59% of the AI agent requests HUMAN Security saw on the sites it protects, and 85% of all agent requests went to product and search pages.

    What happened

    Muse is Meta's personal AI agent. It browses websites for people, compares products and can buy things. To show it's the real Muse and not a scraper using its name, it signs its requests with a digital key. DataDome reports that Muse moved those signatures to a new home, muse.ai, over the weekend of 7 to 9 November 2026. The change is small, but DataDome calls Muse the most widely used agentic browser today, and in HUMAN Security's October data it made 59% of all agent requests, and most of them went to product and search pages.

    What this means for your store

    1. Ask your hosting, CDN or bot-protection provider one question: "Do you recognise Muse's new muse.ai signature?" If the answer is no, Muse may be blocked or slowed down without anyone noticing.
    2. Check your custom rules: if you or a developer added a rule that allows "verified Meta agents", update it so muse.ai is allowed too. Our post on firewalls blocking AI bots shows where these blocks usually hide.
    3. Don't rely on the name in the request: about half of Muse's visits are still unsigned, and anyone can copy a bot name. Decide how you treat agents you can't verify, as in our guide to controlling personal AI agents.
    4. Watch your product pages: that is where agents spend their time. Clear prices, stock and shipping details help an agent recommend you.

    The technical details

    Web Bot Auth uses HTTP Message Signatures (RFC 9421). Each signed request carries a Signature-Agent header that points to the operator's public keys. Muse's header changed from "https://www.meta.com" to "https://muse.ai", and its keys now sit at muse.ai/.well-known/http-message-signatures-directory. DataDome says the two directories overlapped for about 60 hours and Muse's signed traffic stayed level, so the switch was a handover, not a drop. Lists of trusted signers need regular review, because operators add and rotate key directories.

    Want to know if AI agents can reach, read and buy from your site? Check it with Web2Agents.

    FAQ

    Sources

    Newsletter

    New guides on AI visibility, in your inbox.

    Practical, sourced articles on AI crawlers, llms.txt, structured data and AI search. A new one every day.

    Unsubscribe anytime. See our Privacy Policy.