FlowMatrixStudioBot

Who runs it

FlowMatrixStudioBot is operated by FlowMatrix AI Inc., 13 Thacher Street, Boston, MA 02113, which trades as FlowMatrix Studio. We build websites for owner-operated businesses and then run them.

Every request it makes carries this user agent, character for character:

FlowMatrixStudioBot/1.0 (+https://flowmatrixai.studio/bot)

It does not send any other user agent, and it does not disguise itself as an ordinary browser — including on the headless-browser attempt described below.

What it has done so far

On 2 September 2026 this crawler made a one-off measurement pass over about a hundred business websites in Ohio, to see how it behaves — whether it is blocked, how long pages take, how many bytes they carry. Most were visited once. Twelve were fetched and then also rendered, which is the behaviour we have since removed: a refusal now ends it. It read robots.txt first every time and identified itself as it does here. Nobody was contacted as a result, and no sales email has been sent to anyone.

Beyond that it is not yet in regular operation. If you found this name in your log and want to know whether it was us and when, email the address below and we will tell you exactly what we requested and when.

What it asks for

• /robots.txt for the site we are visiting, before anything else on it.
• Publicly available pages, fetched the way a browser fetches them.
• The stylesheets, scripts and images those pages reference, because a page cannot be read without them.

Our own requests use GET and HEAD only. When we render a page, code served by the page may cause the browser to issue additional requests, including POST requests and analytics beacons, exactly as it would in a visitor's browser. We do not choose those requests, but we did start the render that produced them. A single visit is capped at 220 requests and 12 MB, whichever comes first, and stops there.

What it will not do

• Sign in, or request anything behind a login.
• Submit a form — not a contact form, not a quote request, not a newsletter box.
• Go looking for paths of our own — we follow what the page and its scripts reference, and we never probe for admin panels, backups or known vulnerable files.
• Work around a block, a rate limit, or a challenge.
• Follow a redirect from HTTPS to HTTP, or more than three redirects in a row.

How hard it pulls

Our own requests to your site go one at a time, never in parallel, with at least two seconds between them. Where robots.txt sets a longer Crawl-delay, the longer figure wins.

When we render a page, the browser loads that page's own stylesheets, scripts, fonts and images the way any visitor's browser does — in parallel, without our spacing, and including whatever third-party addresses the page itself references. We do not read robots.txt for those third parties, any more than a visitor's browser would. A whole render is capped at 220 requests and 12 MB, counting everything the browser fetched, and stops there.

How to block it

It reads /robots.txt for the site we are visiting before it touches anything else there, and remembers the answer for a day. To refuse it outright:

User-agent: FlowMatrixStudioBot
Disallow: /

If /robots.txt returns a server error, it treats that as a refusal and requests nothing. Three server errors in a row and it stops requesting anything from the site for thirty days.

What happens if you block it

This is the part most crawler pages leave out. A refusal ends it. If robots.txt disallows us, or a firewall, challenge page or error refuses the request, we stop. We do not retry with a headless browser, we do not present a different identity, and we do not deliberately rotate proxies, identities or addresses to get around a block. We record that we were refused and do not come back for thirty days.

We do use a headless browser to render pages, because some pages cannot be read without running their scripts. That is how we read a page we are allowed to read — never a second attempt at one we were refused.

A block and an opt-out are different, and we treat them differently. A technical refusal — a challenge page, a rate limit, a temporary error — puts us off your site for thirty days, after which we may look again, on the same terms, with the same identity. An opt-out is not on a timer. Ask us to stop and we stop, and we do not come back unless you tell us to.

How to opt out entirely

Email [email protected] with the domain you control. A person reads it and replies. There is no form to fill in and no account to create, and we do not ask you to prove anything beyond being able to receive mail at the domain.

An opt-out stops the crawler and stops further unsolicited sales outreach from FlowMatrix Studio about your website — by email, by phone or otherwise. It does not affect messages you ask us for, or service messages if you later become a customer.

Why it exists

We look at how small business websites are built so we can tell an owner what is wrong with theirs, and offer to rebuild it. What the crawler records is what a visitor's browser would have seen: which pages loaded, what they contained, and what the server said. We publish nothing about any site we scan.