IssueMintBot
IssueMintBot fetched a page from your server because somebody using Issue Mint pasted its address into a newsletter draft and we read the page to fill in its title and description for them.
It is not a crawler
IssueMintBot does not spider. It never follows a link it finds in a page, it keeps no index, and it does not come back on a schedule. Every request it makes is one page, requested once, because a person was looking at that address at that moment and asked our composer to describe it. It behaves like the link preview in a chat application rather than like a search engine.
The full user agent string is IssueMintBot/1.0 (+https://issuemint.com/bot).
What it reads, and what it keeps
It asks for HTML and reads four things from it: the canonical URL, the title, the description, and the social preview image. Those go into the draft the publisher is writing, where they will almost certainly edit them. Nothing else is stored, and the page body is discarded once those four values have been taken from it.
How it behaves
- It waits five seconds for a response and then gives up.
- It reads at most one megabyte and stops there.
- It follows at most three redirects, checking each hop before taking it.
- It only speaks HTTP and HTTPS, on ports 80 and 443.
- It sends no cookies and does not run JavaScript.
It does not read robots.txt, and here is why
IssueMintBot does not consult robots.txt. That file governs crawlers
deciding what to visit, and this is not a crawler making that decision: it is a single
fetch of one address a person has just typed, on their behalf. The same is true of the
preview fetchers in Slack, in iMessage and in every social network, none of which
consult it either.
We would rather say that plainly than imply a compliance we have not built. If you think we have it wrong, the address below reaches a person.
Stopping it
Block the user agent at your server, CDN or firewall. Matching on
IssueMintBot is enough and will stay accurate across versions. A publisher
pasting your address will then get an empty title and description and will type their
own, which costs them a few seconds and costs you nothing.
If it is causing you a real problem rather than a theoretical one, write to info@issuemint.com and say so. A request arriving often enough to notice is more likely to be a bug at our end than a publisher pasting the same link repeatedly, and we would want to know.