How to Make Your Website Visible to Meta Muse (GEO for Agentic Browsers)
Meta Muse reaches websites in two ways: through connectors, and by driving a real Chromium browser in a cloud virtual machine. Being visible to Muse therefore means two things. Your pages must be readable by an agent, and your site must work when an agent tries to act on it.
Short answer: make your site easy for a browser-driving agent to read and operate. Keep the markup semantic, labels accessible, layouts stable, prices in plain HTML, and schema accurate. Then decide whether a connector is worth building. Unlike a search crawler, Muse doesn't only read your page: it tries to use it.
How Meta Muse reaches your website
Muse runs in its own cloud virtual machine (Meta calls it Muse Secure VM). When it needs something from a service, it has two paths:
- A connector, if the service has one in the Muse directory. This path is fast and structured.
- A browser. It opens the site in a real Chromium browser and clicks, types and reads like a person. GeekWire reported that Muse does this without announcing itself as an agent.
For most businesses, path two is the reality today. Your website is your API as far as Muse is concerned.
What Meta's crawlers do and don't obey
Meta's crawler documentation lists five agents. None is named Muse.
| User agent | Purpose | Obeys robots.txt? |
|---|---|---|
| Meta-WebIndexer | Improves Meta AI search results | Yes |
| Meta-ExternalAgent | Training foundation models, improving products | Yes |
| Meta-ExternalFetcher | Fetches individual links at a user's request | May bypass |
| Meta-ExternalAds | Advertising and business products | Yes |
| FacebookExternalHit | Link previews when content is shared | Generally; may bypass for security checks |
Two takeaways:
- robots.txt won't reliably keep Muse out. When Amazon wanted Muse gone on September 20, 2026, it used its Conditions of Use and an on-page notice. It couldn't use robots.txt, because there was no Muse user agent to name.
- Don't block the fetchers you want answers from. Blocking training crawlers like Meta-ExternalAgent is a reasonable choice. Blocking Meta-WebIndexer or Meta-ExternalFetcher can remove you from Meta AI answers without stopping browser-based agents.
The Muse-readiness checklist
1. Make the page readable without guessing
- One clear
<h1>, logical headings, and answer-first paragraphs. - Prices, availability, hours and policies in visible HTML text, not only in images or scripts.
- Organization or LocalBusiness JSON-LD on the homepage and Product and Offer schema on product pages. See schema for shopping agents.
- A real
/llms.txttext file that indexes your key pages. Here's how to create one.
2. Make the page operable
- Every button and input needs an accessible name. Agents navigate the accessibility tree, not your visual design.
- Keep layouts stable. Layout shift makes agents click the wrong thing.
- Offer guest checkout and plain form flows. Multi-step modals and hover-only menus are where agents fail.
- Avoid interstitials that hide the content until someone dismisses them.
3. Make the page trustworthy to act on
- State return, shipping and cancellation terms in text near the buy button.
- Keep prices consistent across the page, the schema and checkout. Agents compare them.
- Payment through common wallets helps. Muse pays through Link by Stripe, with Shop Pay and PayPal announced as connectors.
4. Decide on a connector
If customers would delegate a task to an agent (booking, ordering, checking, reporting), a connector gives Muse a direct path that doesn't depend on your markup. Our submission checklist covers the process.
Should you block Muse?
Some large retailers will, as Amazon did. For most small and mid-sized businesses, a customer's agent arriving with intent to buy is a customer. The better move is to make the agent's job easy and keep your terms clear about what automated access is allowed.
Check your site now
Run the free agent readiness check or the AI crawlability checker. For more background, read What Is Meta Muse?, and if you'd rather have us do it, see the AI visibility audit.
Frequently asked questions
How does Meta Muse access websites?
Muse runs in a cloud virtual machine and, where a service has no API or connector, operates websites through a real Chromium-based browser the way a person would. Reporting from GeekWire in September 2026 says it does not identify itself as an agent when it browses.
Can I block Meta Muse with robots.txt?
Not reliably. Meta's crawler documentation does not list a Muse user agent, and Meta-ExternalFetcher, which fetches links at a user's request, may bypass robots.txt. Amazon blocked Muse through its Conditions of Use rather than robots.txt.
Which Meta crawlers should I allow for AI visibility?
Meta documents Meta-WebIndexer for improving Meta AI search results and Meta-ExternalFetcher for user-requested fetches. Meta-ExternalAgent is used for training foundation models. Blocking training crawlers does not stop an assistant fetching your pages when a user asks.
What makes a website easy for Muse to use?
A clean accessibility tree with labelled buttons and form fields, stable layouts with low Cumulative Layout Shift, prices and availability in plain HTML, accurate Organization and Product schema, and a working guest checkout.
Should I build a Muse connector or optimize my website?
Both, if you can. Website optimization helps every agent, not only Muse. A connector gives Muse a reliable, direct path to your service and puts you in its directory.