Privacy policy
Effective 18 August 2026 · applies to the MurmCrawl browser extension, published by MurmTools.
MurmCrawl collects nothing, transmits nothing and stores nothing. It has no server to send data to, no account, no analytics, and no network code of any kind. This page explains exactly how that is true rather than asking you to take it on trust.
What the extension does
When you click the MurmCrawl icon, it asks the site you are already on for its own rules file and shows you one report: which named AI crawlers that file refuses, which it names and allows, and which it never mentions — plus the robots directives on the page itself and in its response header. The report is produced and displayed entirely on your machine, and discarded when you close the panel.
The requests it makes
This is the only MurmTools extension that makes a network request, so here is exactly what it asks for. On a click, and never before one, it requests three things on the origin of the page you are already on — not on any other host:
| Request | Why |
|---|---|
GET /robots.txt | The rules file itself. Read up to 500,000 bytes; if it is longer, the panel says so. |
GET /llms.txt | Reported as present or absent. Nothing is claimed for it — it is a proposed convention no crawler is obliged to honour. |
HEAD on the page you are on | To read its X-Robots-Tag response header, which cannot be seen in the markup. |
All three are sent with credentials: 'omit', so no cookies and no session are
attached. Nothing is sent to us or to anyone else — there is nowhere for it to go, because
this extension has no server. The responses are parsed as text on your machine and thrown away
when the panel closes. Because the header request carries no session, a site that varies its
headers depending on who is signed in may have sent you something different; the panel says so
next to the value rather than hiding it.
Permissions
| Permission | Why it exists |
|---|---|
activeTab |
Grants temporary access to the single tab you clicked the icon on, and only at that moment. It is used to read the figures already published on that page. The access ends with the tab. |
scripting |
Runs one read-only function, once, in that tab. The function is bundled in the extension. It reads the page and requests the three same-origin paths listed above. It does not modify the page, attaches no listeners, and does not persist after returning its result. |
There are no host permissions. MurmCrawl has no access to any website until you click its icon on that website. It cannot run in the background, and it never touches another tab.
Data we collect
None. Collection means transferring data off your device. MurmCrawl has no code path
that sends anything anywhere: no fetch, no XMLHttpRequest, no
WebSocket, no sendBeacon. The only URL in the package is a link to
murmtools.com in the panel footer, which does nothing until you click it.
Data we store
None. There is no storage permission, so the reading exists only while
the popup is open and is gone when you close it. Open the same page twice and it is read from
scratch both times.
Third parties
There are none. No analytics, no error reporting, no A/B testing, no advertising, no payment processor — MurmCrawl is free and has no paid tier.
Even the typefaces are bundled inside the extension rather than fetched from a font host. A tool that reports what a page loads should not quietly load something itself.
What MurmCrawl will not tell you
It does not tell you whether any crawler obeyed. That cannot be known from here or from anywhere. A robots.txt is a request, not a fence: it stops nobody who ignores it, nobody who renames their user agent, and it removes nothing from a model that has already read the site. It also will not tell you whether a site was used for training, because no page carries that fact. Where a figure cannot be known, the panel prints an em-dash and the reason.
Children
MurmCrawl is a publishing and development tool. It is not directed at children and collects no data from anyone, of any age.
Changes
If this policy ever changes, the effective date above changes with it. Since the extension collects nothing, any change would be a narrowing rather than a broadening — and if that ever stopped being true, it would be stated here before it shipped.
Contact
Questions about this policy: [email protected].