Jails for other web servers are obviously as easy.

There was a comment in a different thread that suggested they may respect the robots.txt for the most part, but may ignore wildcards: https://news.ycombinator.com/item?id=46975726

Maybe this is worth trying out first, if you are currently having issues.

If possible block I would block by country first. Even on public websites I block Russia/China by default and that reduced port scans etc.

On "private" services where I or my friends are the only users, I block everything except my country.

Using an LLM to ponder responses for requests is way too costly and slow. Much easier to just use the shotgun approach and fire off a lot of requests and deal with whatever bothers to respond.

This btw is nothing new. Way back when I still used wordpress, it was quite common to see your server logs filling up with bots trying to access endpoints for commonly compromised php thingies. Probably still a thing but I don't spend a lot of time looking at logs. If you run a public server, dealing with maliciously intended but relatively harmless requests like that is just what you have to do. Stuff like that is as old as running stuff on public ports is.

And the offending parties writing sloppy code that barely works is also nothing new.

AI opportunism certainly has added a bit of opportunistic bot and scraper traffic but it doesn't actually change the basic threat model in any fundamental way. Previously version control servers were relatively low value things to scrape. But code just became interesting for LLMs to train on.

Anyway, having any kind of thing responding on any port just invites opportunistic attempts to poke around. Anything that can be abused for DOS purposes might get abused for exactly that. If you don't like that, don't run stuff on public servers or protect them properly. Yes this is annoying and not necessarily easy. Cloud based services exist that take some of that pain away.

Logs filling up with 404, 401, or 400 responses should not kill your server. You might want to implement some logic that tells repeat offenders 429 (too many requests). A bit heavy handed but why not. But if you are going to run something that could be used to DOS your server, don't be surprised if somebody does that.

I think it’s a) volume of scrapers, and b) desire for _all_ content instead of particular content, and c) the scrapers are new and don’t have the decades of patches Googlebot et al do.

5 years ago there were few people with an active interest in scraping ForgeJo instances and personal blogs. Now there are a bajillion companies and individuals getting data to train a model or throw in RAG or whatever.

Having a better scraper means more data, which means a better model (handwavily) so it’s a competitive advantage. And writing a good, well-behaved distributed scraper is non-trivial.

I'm hazarding a guess that there are many AI startups that focus on building datasets with the aim to sell those datasets. Still doesn't make total sense, since doing it badly would only hurt them, but maybe they don't really care about the product / outcome, they're just capturing their bit of the AI goldrush?

> why they're attributed to AI?

I don’t think they mean scrapers necessarily driven by LLMs, but scrapers collecting data to train LLMs.

I’m guessing, but I think a big portion of AI requests now come from agents pulling data specifically to answer a user’s question. I don’t think that data is collected mainly for training now but are mostly retrieved and fed into LLMs so they can generate the response. Thus so many repeated requests.

I Did not know this! Thanks man!

https://anubis.techaro.lol/docs/admin/honeypot/overview The Anubis scraper protection has this as a feature. Just sends garbage if something falls into a trap.

> this is pretty much the reason why Cloudflare even exists,

You said it yourself. If you're selling a cure, you might as well start a plague.

I just threw up a public Forjego instance for some lightweight collaboration. About 2 minutes after the certificate was created, I'm guessing they picked up the instance from the transparency logs for certificates, and started going through every commit and so on from the two repositories I had added.

Watched it for a while, thinking eventually it'd end. It didn't, seemed like Claudebot and GPTBot (which was the only two I saw, but could have been forged) went over the same URLs over and over again. They tried a bunch of search queries too at the same time.

The day after I got tired of seeing it so added a robot.txt forbidding any indexing. Waited a few hours, saw that they were still doing the same thing, so threw up basic authentication with `wiki:wiki` as the username:password basically, wrote the credentials on the page where I linked it and as expected they stopped trying after that.

They don't seem to try to bypass anything, whatever you put in front will basically defeat them except blocking them by user-agent, then they just switch to a browser-like user-agent instead, which is why I went the "trivial basic authentication" path instead.

Wasn't really an issue, just annoying when they try to masquerade as normal users. Had the same issue with a wiki instance, added rate limits and eventually they seemingly backed off more than my limits were set too, so I guess they eventually got it. Just checked the logs and seems they've stopped trying completely.

Seemingly it seems like people who are paying for their hosting by usage (which never made sense to me) is the ones hard hit by this. I'm hosting my stuff on a VPS, and don't understand what the big issue is, worst case scenario I'd add more aggressive caching and it wouldn't be an issue anymore.

I would love to understand this.

Just a few years ago badly behaved scrapers were rare enough not to be worth worrying about. Today they are such a menace that hooking any dynamic site up to a pay-to-scale hosting platform like Vercel or Cloud Run can trigger terrifying bills on very short notice.

"It's for AI" feels like lazy reasoning for me... but what IS it for?

One guess: maybe there's enough of a market now for buying freshly updated scrapes of the web that it's worth a bunch of chancers running a scrape. But who are the customers?

> Does anyone know what's the deal with these scrapers, or why they're attributed to AI?

You don't really need to guess, it's obvious from the access logs. I realize not everyone runs their own server, so here are a couple excerpts from mine to illustrate:

- "meta-externalagent/1.1 +https://developers.facebook.com/docs/sharing/webmasters/craw...)"

- "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)"

- "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Amazonbot/0.1; +https://developer.amazon.com/support/amazonbot) Chrome/119.0.6045.214 Safari/537.36"

- "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; GPTBot/1.3; +https://openai.com/gptbot)"

- [...] (compatible; PetalBot;+https://webmaster.petalsearch.com/site/petalbot)"

And to give a sense of scale, my cgit instance recieved 37 212 377 requests over the last 60 days, >99% of which are bots. The access.log from nginx grew to 12 GiB in those 60 days. They scrape everything they can find, indiscriminately, including endpoints that have to do quite a bit of work, leading to a baseline 30-50% CPU utilization on that server right now.

Oh, and of course, almost nothing of what they are scraping actually changed in the last 60 days, it's literally just a pointless waste of compute and bandwidth. I'm actually surprised that the hosting companies haven't blocked all of them yet, this has to increase their energy bills substantially.

Some bots also seem better behaved then others, OpenAI alone accounts for 26 million of those 37 million requests.

I remember back when Anubis came out, some naysayers on here were saying it wouldn't work for long because the scrapers would adapt. Turns out careless, unethical vibecoders aren't very competent.

You know, I reckon if you serve up smut or instructions on bomb creation or something they stop hammering you...

> I set it and use JavaScript to tell them to reload the page

While throwing out all users who opt-in to javascript, using Noscript or uBlock or something like it, may be acceptable collateral damage to you, it might be good to keep in mind that this plays right into Big Adtech's playbook. They spend over two decades to normalize the behavior of running a hundred or more programs of untrusted origin on every page load, and to treat users to opt-in to running code in a document browser with suspicion. Not everyone would like to hand over that power to them on a silver platter with a neat little bow on top.

Alternatively, from the nginx config file for git.ardour.org:

   location ~ commit/* {
        return 404;
    }

I got another interesting idea from this and another comment but what if we combine this with ssh git clients/websites with the normal ability.

maybe something like https://ssheasy.com/ or similar could also be used? or maybe even a gotty/xterm instance which could automatically ssh/get a tui like interface.

I feel as if this would for all scrapers be enough?

Using gitea does not help if you goal is to allow non-auth'ed read-only access to the repo from a web browser. The scrapers use that to hit up every individual commit, over and over and over.

We used nginx config to prevent access to individual commits, while still leaving the "rest" of what gitea makes available read-only for non-auth'ed access unaffected.

The dirty secret is a lot of them come through "residential proxies", aka backdoored home routers, iot devices with shitty security, etc. Basically the scrapers who are often also third party, go to these "companies" and buy access to these "residential proxies". Some are more... considerate than others.

Why? Data. Every bit of it is it might be valuable. And not to sound tin foil hatty, but we are getting closer to a post-quantum time (if we aren't already ).

I doubt it's OpenAI. Maaaybe somebody who sells to OpenAI, but probably not. I think they're big enough to do this mostly in-house and properly. Before AI only big players would want a scrape of the entire internet, they could write quality bots, cooperate, behave themselves, etc. Now every 3rd tier lab wants that data and a billion startups want to sell it, so it's a wild west of bad behavior and bad implementations. They do use residential IP sets as well.

The post says their web frontend was public.

Just block the whole China, India and similar countries.

Start blocking /16s.

I've put everything behind a Wireguard Server, so if I need something, I can access to it through VPN and AI can't do anything

Cloudflare launched a product to do that last year: https://blog.cloudflare.com/introducing-pay-per-crawl/

I have no idea if it actually works as advertised though. I don't think I've heard from anyone trying it.

A healthy front page shouldn’t be a “famous people only” section. If only big names can show up there, it’s not discovery anymore, it’s just a popularity scoreboard.

Well the fact that this supposed nobody is overwhelmed by AI scrapers should speak a lot about the issue no?

Send it where?

the era of mourning has begun

you don't understand what self-hosting means. self-hosting means the site is still up when AWS and Cloudflare go down.

There's value to be had in ripping the copyright off your stuff so someone else can pass it off as their stuff. LLMs have no technical improvements so all they can do is throw more and more stolen data into it and hope it, somehow, crosses a nebulous "threshold" where it suddenly becomes actually profitable to use and sell.

It's a race to the bottom. What's different is we're much closer to the bottom now.

I stopped trying to understand. Encountering a 404 on my site leads directly to a 1 year ban.

The big nasty AI bots use 10s of thousands of IPs distributed all over China

A lot of compromised home devices and cheap servers proxying traffic, from all over the world.

I'm also dealing with a scraper flood on a cgit instance. These conclusions come from just under 4M lines of logs collected in a 24h period.

- Caching helps, but is nowhere near a complete solution. Of the 4M requests I've observed 1.5M unique paths, which still overloads my server.

- Limiting request time might work, but is more likely to just cause issues for legitimate visitors. 5ms is not a lot for cgit, but with a higher limit you are unlikely to keep up with the flood of requests.

- IP ratelimiting is useless. I've observed 2M unique IPs, and the top one from the botnet only made 400 well-spaced-out requests.

- GeoIP blocking does wonders - just 5 countries (VN, US, BR, BD, IN) are responsible for 50% of all requests. Unfortunately, this also causes problems for legitimate users.

- User-Agent blocking can catch some odd requests, but I haven't been able to make much use of it besides adding a few static rules. Maybe it could do more with TLS request fingerprinting, but that doesn't seem trivial to set up on nginx.

Great, now we need caching for something that's seldom (relatively speaking) used by people.

Let's not forget that scrapers can be quite stupid. For example, if you have phpBB installed, which by defaults puts session ID as query parameter if cookies are disabled, many scrapers will scrape every URL numerous times, with a different session ID. Cache also doesn't help you here, since URLs are unique per visitor.

because he's unable to self-host git anymore because AI bots are hammering it to submit PRs.

self-hosting was originally a "right" we had upon gaining access to the internet in the 90s, it was the main point of the hyper text transfer protocol.

I would say there's a couple aspects.

The crawlers for the big famous names in AI are all less well behaved and more voracious than say, Googlebot. Though this is all somewhat muddied by companies that ran the former "good" crawlers all also being in the AI business and sometimes trying to piggyback on people having allowed or whitelisted their search crawling User-Agent, mostly this has settled a little where they're separating Googlebot from GoogleOther, facebookexternalhit from meta-externalagent, etc. This was an earlier "wave" of increased crawling that was obviously attributable to AI development. In some cases it's still problematic but this is generally more manageable.

The other stuff, the ones that are using every User-Agent under the sun and a zillion datacenter IPs and residential IPs and rotate their requests constantly so all your naive and formerly-ok rate-based blocking is useless... that stuff is definitely being tagged as "for AI" on the basis of circumstantial evidence. But from the timing of when it seemed to start, the amount of traffic and addresses, I don't have any problem guessing with pretty high confidence that this is AI. To your question of "who are the customers"... who's got all the money in the world sloshing around at their fingertips and could use a whole bunch of scraped pages about ~everything? Call it lazy reasoning if you'd like.

How much this traces back ultimately to the big familiar brand names vs. would-be upstarts, I don't know. But a lot of sites are blocking their crawlers that admit who they are, so would I be surprised to see that they're also paying some shady subcontractors for scrapes and don't particularly care about the methods? Not really.

Huh, I had a gitea instance in the public web on one of my netcup vps's. I didn't set any logs and was using cloudflare tunnels (with a custom bash script which makes cf tunnels expose PORT SUBDOMAIN).

Maybe its time for me to go ahead and start it again with logs to see if there are any logs.

I will maybe test it in all three 1) With CF tunnels + AI Block, 2) Only CF tunnels, 3) On a static IP directly. Maybe you can try the experiment too and we can compare our findings (also saying because I am lazy and I had misconfigured that cf tunnel so when it quit, I was too lazy to restart the vps given I just use it as a playground and just wanted to play around self hosting but maybe I will do it again now)

The bar to ingest unstructured data into something usable was lowered, causing more people to start doing it.

Used to be you needed to implement some papers to do sentiment analysis. Reasonably high bar to entry. Now anyone can do it, the result: more people doing scraping (in less competent scrapers too).

> Turns out careless, unethical vibecoders aren't very competent.

Well they are scraping web pages from a git forge, where they could just, you know, clone the repo(s) instead.

I think that they actually do. I remember either some discussion (so a HN post) or a HN comment actually talking about it. Oh I should've favourited it but yes this (sort of) actually works (Maybe someone can test it?)

"Turns out careless, unethical vibecoders aren't very competent." well, they rely on AI, don't they? and AI is trained with already existing bad code, so why should the outcome be different?

Lately it seems like every time I start reading through a comment with a bunch of incoherent word and idea salad, I look up and there's your username.

i'm working on something similar: instead of web-based ssh client, it's a web-based git client UI - you can "checkout" repos, browse commits, tree, read individual files, etc. with no server-side code at all; git objects are fetched and parsed on client-side. first target is the dumb-http git protocol, so people can host git repos on static websites, and visitors don´t need to clone by a local git client to peek in.

https://bandie91.github.io/dumb-http-git-browser-js-app/ui.h...

it isn't that hard to just buy a bunch of sim cards and put them in a modem and use that. it's good enough as a residential proxy. source: I did that before, when I worked on plaid-like thing.

Stop just making up excuses for these companies. Other comments on this story have showed the bots are using openai user agents and making requests from openai owned ip ranges.

Since you had the logs for this, can you confirm the IP ranges they were operating from? You mention "Claudebot and GPTBot" but I'm guessing this is based off of the user-agent presented by the scrapers and could easily be faked to shift blame. I genuinely doubt Anthropic and such would be running scrapers that are this badly written/implemented, it doesnt make economic sense. I'd love to see some of the web logs from this if you'd be willing to share! I feel like this is just some of the old scraper bots now advertising themselves as AI bots to shift blame into the AI companies.

I had the same issue when I first put up my gitea instance. The bots found the domain through cert registration in minutes, before there were any backlinks. GPTbot, ClaudeBot, PerplexityBot, and others.

I added a robots.txt with explicit UAs for known scrapers (they seem to ignore wildcards), and after a few days the traffic died down completely and I've had no problem since.

Git frontends are basically a tarpit so are uniquely vulnerable to this, but I wonder if these folks actually tried a good robots.txt? I know it's wrong that they ignore wildcards, but it does seem to solve the issue

For whatever reason, legislation is lax right now if you claim the purpose of scraping is for AI training even for copyrighted material.

May be everyone is trying to take advantage of the situation before law eventually catches up.

Following your link above, https://openai.com/gptbot

> ChatGPT-User is not used for crawling the web in an automatic fashion. Because these actions are initiated by a user, robots.txt rules may not apply.

So, not AI training in this case, nor any other large-batch scraping, but rather inference-time Retrieval Augmented Generation, with the "retrieval" happening over the web?

I still think it is just a matter of time until scrapers catch up. There are more and more scrapers that spin up an full blown chromium.

Oh please. That ship has sailed. I'm marginally sympathetic to people who don't run JavaScript on their browsers for a variety of reasons, but they've deliberately opted out of the de facto modern web. JS is as fundamental to current design as CSS. If you turn it off, things might work, but almost no one is testing that setup, nor should they reasonably be expected to.

This has zero to do with Adtech for 99.99% of uses, either. Web devs like to write TypeScript and React because that's a very pleasant tech stack for writing web apps, and it's not worth the effort for them to support a deliberately hamstrung browser for < 0.1% of users (according to a recent Google report).

See also: feel free to disable PNG rendering, but I'm not going to lift a finger to convert everything to GIFs.

Every commit. Every diff between 2 different commits. Every diff with different query parameters. Git blame for each line of each commit.

Imagine a task to enumerate every possible read-only command you could make against a Git repo, and then imagine a farm of scrapers running exactly one of them per IP address.

Ugh.

Has this actually been investigated and proven to be true? I see allegations, but no facts really.

It seems to me to be just as likely that people are installing LLM chatbot apps that do the occasional bit of scraping work on the sly, covered by some agreed EULA.

Yuk…

   http {
       # ... other http settings
       limit_req_zone $binary_remote_addr zone=mylimit:10m rate=10r/s;
       # ...
   }


    server {
        # ... other server settings
        location / {
            limit_req zone=mylimit burst=20 nodelay;
            # ... proxy_pass or other location-specific settings
        }
    }

Rate limit read-only access at the very least. I know this is a hard problem for open source projects that have relied on web access like this for a while. Anubis?

How can I detect if my router is backdoored, or being used as a residential proxy?

Millions and all over the world

You’re describing changing the base assumption for software reachable on the internet. “Assume all possible unauthenticated urls will be hit basically constantly”. Bots used to exist but they were rare traffic spikes that would usually behave well and could mostly be ignored. No longer.

So... just blacklist all China IPs? I assume China isn't the primary market for most of complaining site-owners.

Cloudflare seems to be taking over all of the last mile web traffic, and this extreme centralization sounds really bad to me.

We should be able to achieve close to the same results with some configuration changes.

AWS / Azure / Cloudflare total centralization means no one will be able to self host anything, which is exactly the point of this post.

Cloudflare won't save you from this - see my comment here: https://news.ycombinator.com/item?id=46969751#46970522

For logging, statistics etc. we have the Cloudflare bot protection on the standard paid level, ignore all IPs not from Europe (rough geolocation), and still have over twice the amount of bots that we had ~2 years ago.

I'm pretty sure scrapers aren't supposed to act as low key DOS attacks

I think the point of the post was how something useless (AI) and its poorly implemented scrapers is wrecking havoc in a way that’s turning the internet into a digital desert.

That Cloudflare is trying to monetise “protection from AI” is just another grift in the sense that they can’t help themselves as a corp.

Sounds like you're keeping all your URLs alive forever? Commendable!

Damn, as someone who sometimes navigate by guessing URLs and rewriting them manually in the address bar, I hope more don't start doing this, I probably see at least one self-inflicted 404 per day at least.

If that is the case how can you determine the reason for the activity?

They're rotating through huge pools of residential IP addresses.

The scrapers should use some discretion. There are some rather obvious optimizations. Content that is not changing is less likely to change in the future.

They don't. I'm using Cloudflare and 90%+ of the traffic I'm getting are still broken scrapers, a lot of them coming through residential proxies. I don't know what they block, but they're not very good at that. Or, to be more fair: I think the scrapers have gotten really good at what they do because there's real money to be made.

Cool, I can take all my self hosted stuff and stick it behind centralised enterprise tech to solve a problem caused by enterprise tech. Why even bother?

There are a bit too many IPs to list but from my logs they're always of the form 74.7.2XX.* for GPTBot, matching OpenAIs published ip ranges[0].

So yes, they are definitely running scrapers that are this badly written.

Also old scraper bots trying to disguise themselves as GPTBot seems wholly unproductive, they're try to immitate users, not bots.

[0] https://openai.com/gptbot.json

> but I'm guessing this is based off of the user-agent presented by the scrapers and could easily be faked to shift blame

Yes, hence the "which was the only two I saw, but could have been forged".

> I'd love to see some of the web logs from this if you'd be willing to share!

Unfortunately not, I'm deleting any logs from the server after one hour, and also don't even log the full IP. I took a look now and none of the logs that still exists are from any user agent that looks like one of those bots.

I will second a good robots.txt. Just checked my metrics and < 100 requests total to my git instance in the last 48 hours. Completely public, most repos are behind a login but there are a couple that are public and linked.

> For whatever reason, legislation is lax right now if you claim the purpose of scraping is for AI training even for copyrighted material

I think the reason is that America & China for the most part are also in AI arms race combined with an AI bubble and neither side would wish to lose literally any percieved advantage to them no matter the cost on others.

Also there is an immense lobbying effort against senators who propose for a stricter AI regulation.

https://www.youtube.com/watch?v=DUfSl2fZ_E8 [What OpenAI doesn't want you to know]

It's actually a great watch. Highly recommended because a lot of talks about regulations does feel to me as mirrors and smoke.

Likely, at least for some. I've caught various chatbots/CLI harnesses more than once inspecting a github repo file by file (often multiple times, because context rot)

But the sheer volume makes it unlikely that's the only reason. It's not like everybody has constantly questions bout the same tiny website.

It seems inevitable, but in the mean time, that's vastly more expensive than running curl in a loop. In fact, it may be expensive enough that it cuts bot traffic down to a level I no longer care about defending against. Like GoogleBot had been crawling my stuff for years without breaking the site. If every bot were like that, I wouldn't care.

Even that functions as a sort of proof of work, requiring a commitment of compute resources that is table stakes for individual users but multiplies the cost of making millions of requests.

AFAIK you can bypass it with curl because there's an explicit whitelist for it, no need for a headful browser.

Well it's a race, just like security. And as long as anubis is in the front, all looks bright

> JS is as fundamental to current design as CSS.

I think this hits the crux of the trend fairly well.

And is why I have so many workarounds to shitty JS in my user files.

Because I can't see your CSS, either.

Ugh Ugh Ugh ... and endless ughs, when all they needed was "git clone" to get the whole thing and spend as much time and energy as they wanted analyzing it.

Another likely source is "free" VPN tools, or tools for streaming TV (especially football or other pay-to-view stuff). The tool can make a little money proxying requests at the same time.

I can't provide evidence as it's close to impossible to separate the AI bots using residential proxies from actual users, and their IPs are considered personal data. But as the other reply shows, it's easy enough to find people selling this service.

I'm dealing with such attack, so if you'd like, you can send me IPv4 addresses, and I'll grep my logs for them. Email address is on the website linked on my profile.

As for what you can do on your own, it really depends on your network. OpenWRT routers can run tcpdump, so you can check for suspicious connections or DNS requests, but it gets really hard to tell if you have lots of cloud-tethered devices at home. IoT, browser extensions, and smartphone applications are the usual suspects.

The most surefire way would be to put a device between your router and your ONT/modem to capture the packets and see what requests are being sent. It's not complicated but it IS a lot of information to sift through.

Your router may have the ability to log requests, but many don't, and even if yours does, if you're concerned the device may be compromised, how can you trust the logs?

BUT, with all that said, these attacks are typically not very sophisticated. Most of the time they're searching for routers at 192.168.1.1 with admin/admin as the login credentials. If you have anything else set, you're probably good from 97% of attackers (This number is entirely made up, but seriously that percentage is high). You can also check for security advisories on your model of router. If you find anything that allows remote access, assume you're compromised.

---

As a final note, it's more likely these days that the devices running these bots are IoT devices and web browsers with malicious javascript running.

> How can I detect if my router is backdoored, or being used as a residential proxy?

Aside from the obvious smoke tests (are settings changing without your knowledge? Does your router expose access logs you can check?), I'm not sure there's any general purpose way to check, but 2 things you can do are:

1. search for your router's model number to see if it's known to be vulnerable, and replace it with a brand-new reputable one if so (and don't buy it from Amazon).

2. There are vendors out there selling "residential proxy IP databases", (e.g., [1]) no idea how good they are, but if you have a stable public IP address you could check whether you're on that.

[1] https://ipinfo.io/data/residential-proxy

Where does one find a good robots.txt? Are there any well maintained out there?

> I wonder if these folks actually tried a good robots.txt?

I suspect that some of these folks are not interested in a proper solution. Being able to vaguely claim that the AI boogeyman is oppressing us has turned into quite the pastime.

the recent google report claimed that less than 0.1% of users have javascript disabled ... like for every website, or just some, or?

your PNG/GIF thing is nonsense (false equivalence, at least) and seems like deliberate attempt to insult

> I'm marginally sympathetic

you say that as if they've done some harm to you or anyone else. outside of these three words, you actually seem to see anyone doing this as completely invalid and that the correct course of action is to act like they don't exist.

We used fail2ban to do rate limiting first. It wasn't adequate.

Seriously, go to Google.

Search for: "residential proxy" ai data scraping.

Start reading through thousands of articles.

If it’s legit you can ask your ISP if they sell use of your hardware. Or just don’t use the provided hardware and instead BYO router or modem or media converter or whatever.

But I think what OP is implying is insecure hardware being infected by malware and access to that hardware sold as a service to disreputable actors. For that buy a good quality router and keep it up to date.

Also converting the blog from something dynamic to a static site generator. I made the same switch partly for ease of maintenance, but a side benefit is it's more resilient to this horrible modern era of scrapers far outnumbering legitimate traffic.

It's painful to have your site offline because a scraper has channeled itself 17,000 layers deep through tag links (which are set to nofollow, and ignored in robots.txt, but the scraper doesn't care). And it's especially annoying when that happens on a daily basis.

Not everyone wants to put their site behind Cloudflare.

The 2GB RAM didn't fill up with banned addresses, but YMMV.

Probably more money in scraping than protection...

They don't care. It's the reason they ignore robots.txt and change up their useragents when you specifically block them.

"Cause a problem and then sell the solution" proves a winning business strategy once more.

Quick question but do these bots which you mention are from a 24H period but how long will this "attack" continue for?

Because this is something which is happening continuously & i have observed so many HN posts like these (Anubis iirc was created by its creator out of such frustration too). Git servers being scraped to the point of its effectively an DDOS.

sorry if i missed it, but the original post doesn't say anything about PRs... the bots only seem to be scraping the content

Why would you do that?

Not so many...

And there are tools to scan for dead links.

Some fake user agent, some tell you who they are. Or.. do they?

Here-in is the problem. And if you block them, you risk blocking actual customers.

Cloudflare actually has this as a free tier feature so even if you don't want to use it for your site you can just setup a throwaway domain on Cloudflare and periodically copy the robots.txt they generate from your scraper allow/block preferences, since they'll be keeping up to date with all the latest.

> Being able to vaguely claim that the AI boogeyman is oppressing us has turned into quite the pastime.

FWIW, you're literally in a comment thread where GP (me!) says "don't understand what the big issue is"...

Ooof, maybe a write up is in order? An opinioned blog post? I'd love to know more.

It would be literally impossible to know whether a user disabled JavaScript on another site, so I'm going to say that they meant that for their own sites.

> you say that as if they've done some harm to you or anyone else.

I was literally responding to someone referring to themselves as "collateral damage" and saying I'm playing into "Big Adtech's playbook". I explained why they're wrong.

> the correct course of action is to act like they don't exist.

Unless someone is making a site that explicitly targets users unwilling or unable to execute JavaScript, like an alternative browser that disables it by default or such, mathematically, yes, that's the correct course of action.

That's the worst thing I've seen all week. The DDoS networks of 20 years ago, now out in the open and presented as real business.

Thanks for the info, wish I didn't know :-(

Faster. Wanna know the pricing? $domain/pricing. What's this company about? $domain/about. Switch to another Google account? Change the 1 to a 2 in the URL. I guess mostly to avoid the mouse ultimately.

oh you're right, I read "pointless requests" as "PRs", oops!

So you don't know? Why respond?

Yes, the attack is continuous. The rate fluctuates a lot, even within a day. It's definitely an anomaly, because eg. from 2025-08-15 to 2025-10-05 I saw zero days with more than 10k requests. Here's a histogram of the past 2 weeks plus today.

  2026-01-28     21'460
  2026-01-29     27'770
  2026-01-30     53'886
  2026-01-31    100'114  #
  2026-02-01    132'460  #
  2026-02-02     73'933
  2026-02-03    540'176  #####
  2026-02-04    999'464  #########
  2026-02-05    134'144  #
  2026-02-06  1'432'538  ##############
  2026-02-07  3'864'825  ######################################
  2026-02-08  3'732'272  #####################################
  2026-02-09  2'088'240  ####################
  2026-02-10    573'111  #####
  2026-02-11  1'804'222  ##################

If they are using appropriated hardware, what possible reason could there be for them saying who they are?

Can you scan my bookmarks? :) edit: i.e. if someone has a bookmark to a page on your site and it goes 404, then they are blocked for a year. You have no ability to scan it because it's a file on their local system.

Don't be rude.

Three different "companies" normally:

1. The residential proxies

2. Scrapers, on behalf of or as an agent of the data buyer

3. Data buyer (ai training)

Scrapers are buying from residential proxies, giving the data buyer a bit of a shield/deniability.

The scrapers don't want to get outright blocked if they can avoid it, otherwise they have nothing to sell.

Thanks for sharing the data, This unpredictability and everything is even more suspicious.

Thoughts on having an ssh server with https://github.com/charmbracelet/soft-serve instead?

Oh, now I understand.

I never removed anything, but I'll keep this in mind for the future.

It's plausible that the AI companies have given up storing data for training runs and just stream it off the Internet directly now. It's probably cheaper to stream than buying more SSDs and HDDs from a supply constrained supply chain at this point.