I just put Anubis in front of my self-hosted forge this morning because AmazonBot had helped itself to 750 GiB (!) of traffic to my public repos this month!
> We are writing to inform you that starting Monday, June 15, 2026, crawl preferences for Amazonbot will be managed solely through the industry-standard directives.
I hope you leave it on the WAF. If they're only just deciding to respect robots.txt, which has been internet infrastructure forever, then it's probably still incredibly amateur software with 'Amazon-priorities' rather than 'responsible internet traffic' priorities.
Good place to ask, saw a new AWS User agent in logs today: Amazon-Quick-on-Behalf-of-$HEXID
I found a mention on some user agent trackers but no official documentation. Anyone knows if it’s documented? Asking because I am seeing decent traffic (30GB/week) from this.
> Crawling behavior [...] Crawler identification: Identifies itself with user-agent string "aws-quick-on-behalf-of-<UUID>" in request headers.
Maybe people found a way of using it as a loophole for something or Amazon Quick is just picking up in usage, and your website is popular amongst whoever uses that sort of stuff.
> Amazonbot is used to improve our products and services. This helps us provide more accurate information to customers and may be used to train Amazon AI models.
I was wondering about this. And it makes me think this is all mistruth, unless they plan to drop this pricing tactic.
They've been getting some heat on it lately, but I find it hard to believe they're going to give up entirely? And if so, what's to stop someone from just flouting their rules on pricing, and then doing the robots.txt thing to prevent issues?
Amazonbot is specifically the user agent they use for crawling for "provide more accurate information to customers" (whatever that means, could be anything it sounds like) and also when they scrape for data used in AI training, according to https://developer.amazon.com/amazonbot
Huh, I get a lot of traffic from Amazonbot (relative to humans) and try as I might, it would get stuck in a tarpit of no creation because it would sit there and keep blasting every variation of my recent pages because Mediawiki lists many links. I have them appropriately nofollow and warning the bot not to waste its time with robots.txt but it just goes and sticks itself on nonsense internal pages.
The traffic isn't a problem. I've got Cloudflare in front and the machine itself is relatively overpowered, and downtime isn't critical. But I'd just like the thing to be able to spider me properly. Someone did point out to me that maybe I wasn't receiving actual Amazonbot but some other spider: https://news.ycombinator.com/item?id=46352723
Robots.txt is lame BTW, there is no way to enforce it. It is up to the bot to decide to crawl or not and most cases they don't care.
Cloudflare had a nice technic to address the bot problem (if you use their name servers). It'll respect and use the robots.txt while sending the remaining bots to a deep black hole.
Yes, we know, its purpose is to guide the bots, not forcibly block them.
That said, one of the biggest websites in the world not respecting it is definitely a noteworthy story. Hopefully another one of the biggest websites in the world (formerly known as Twitter) eventually respects it as well instead of not even disclosing itself via a user agent and pretending to be Safari running on iOS.
At least, it claimed to be AmazonBot…
They will in the future, but not today.
Did end up just adding them to our WAF blocklist, which is weirdly ironic - hosting on their infra & using their services to block their AI scraper...
this bit made me laugh. was the email drafted in Outlook? was it sent to some sort of forwarding mailbox, or did they just BCC every customer in?
I found a mention on some user agent trackers but no official documentation. Anyone knows if it’s documented? Asking because I am seeing decent traffic (30GB/week) from this.
> Crawling behavior [...] Crawler identification: Identifies itself with user-agent string "aws-quick-on-behalf-of-<UUID>" in request headers.
Maybe people found a way of using it as a loophole for something or Amazon Quick is just picking up in usage, and your website is popular amongst whoever uses that sort of stuff.
It has AI agents included so I guess this can just come from it searching the web based on user requests.
> Amazonbot is used to improve our products and services. This helps us provide more accurate information to customers and may be used to train Amazon AI models.
They've been getting some heat on it lately, but I find it hard to believe they're going to give up entirely? And if so, what's to stop someone from just flouting their rules on pricing, and then doing the robots.txt thing to prevent issues?
The traffic isn't a problem. I've got Cloudflare in front and the machine itself is relatively overpowered, and downtime isn't critical. But I'd just like the thing to be able to spider me properly. Someone did point out to me that maybe I wasn't receiving actual Amazonbot but some other spider: https://news.ycombinator.com/item?id=46352723
Cloudflare had a nice technic to address the bot problem (if you use their name servers). It'll respect and use the robots.txt while sending the remaining bots to a deep black hole.
That said, one of the biggest websites in the world not respecting it is definitely a noteworthy story. Hopefully another one of the biggest websites in the world (formerly known as Twitter) eventually respects it as well instead of not even disclosing itself via a user agent and pretending to be Safari running on iOS.
You're talking about one (yes, biggest) but millions of other bots don't follow must be a bigger story.