MangaDex Development Update - API Abuse Enforcement

[BestBoy]

I guarantee it will be up for longer, though, just like it was back in January.

 

[Tscheddov]

and people called me an idiot in the past when i pointed out MD is going in a bad direction.

Youtube for Manga, eh ?

 

[asbun123]

and people called me an idiot in the past when i pointed out MD is going in a bad direction.

Youtube for Manga, eh ?

Name these people. It's pretty clear that this site is experiencing enshittification

 

[Tscheddov]

Name these people. It's pretty clear that this site is experiencing enshittification

nah, they either see it themselfes now or calling them out would be pointless anyway because they actively dont want to acknowledge it.

i am not here to accuse people personally anyway. i still have not given up on Mangadex.
i still believe a unified userbase can still save this place. without the users, uploaders and groups this Place would be nothing.
the groups, uploders and users on the other hand need this platform too.



i wish i was just a shizophrenic tinfoil-hat. i would prefer being a shizo who is seeing things that arent actually true over what is happening right now.

 

[finalm]

i still believe a unified userbase can still save this place. without the users, uploaders and groups this Place would be nothing.
the groups, uploders and users on the other hand need this platform too.

Clearly they don't, since the majority of groups no longer upload here and many have taken down what content they did in the past.

 

[manadrop]

I like my APIs as I pour my tea: 418

 

[jvlg]

(Disregard any of that if you saw it, I'm dumb and can't read dates. Catching up on the rest of the combo right now)
Yeah, I've got nothing to add that hadn't been said already or turned out to be irrelevant

 

[Gwmd]

I'm experiencing site issues that make me worry this enforcement goes beyond just excessive API calls; I've historically managed manga updates by organizing bookmark folders, since not everything's available all on one site, and once very couple weeks choosing the 'open all in tabs' option to pop fortyish followed series up at once and check through all of 'em for new chapters.

Now though, that seems to be triggering service timeout and hourslong blacklisting; they're all legitimate requests going through a consumer-standard browser, but because there's no apparent attempt to differentiate between 'this client made 40 requests in a second' and 'this client is making 40 requests in a second, every second, for a period indicative of scraping', suddenly junk's way less usable. Like, I'm already authenticating with a validated account and going through cloudflare half a dozen times a day too, how enshittified can the internet get???

E; I can't even access these forums or get a post through reliably - ctrl+F5 loads whatever page I'm targeting, but the moment I try to move to another page element it's back to 404ing me until I ctrl+f5 again - which obviously provides no avenue for using the post form, I had to log into and use an entirely different machine / browser since not even clearing cookies got me back in good graces.

 

[wiz_gmc]

I've historically managed manga updates by organizing bookmark folders, since not everything's available all on one site, and once very couple weeks choosing the 'open all in tabs' option to pop fortyish followed series up at once and check through all of 'em for new chapters.

Now though, that seems to be triggering service timeout and hourslong blacklisting;

It actually takes a lot more than 40 requests to open 40 manga pages. Considering the limit is 5 per second, you have probably already made too many requests before the pages even have a chance to catch themselves doing it.

Assuming you don't want to make any behavioural changes, I made a script that will let you keep using tabs with less requests per second. It makes a waiting room for you, so that your browser doesn't actually load anything from MangaDex until you are looking at the tab.

It will run off a local page so you don't need a server or anything to have it work. But if you did want it to run it off a server, any hosting will do. It will run from a Github pages.

Save this code as an html file.

<!DOCTYPE html><html lang=""><head><meta charset="utf-8"><title></title><script>
const path = document.location.hash.slice(1);
document.title = path.split('/')[2];
addEventListener("focus", (e) => {
    if(!path) {return;}
    window.location.replace(`https://mangadex.org/${path}`);
});
</script></head><body></body></html>

Then edit your bookmarks so that

https://mangadex.org/title/a4807f27-d96e-4069-9705-88056e81a555/ririmu-horikku

becomes

file://filename.html#title/a4807f27-d96e-4069-9705-88056e81a555/ririmu-horikku

It won't be literally file://filename.html, it will be whatever path you used to save the script.

It works perfectly for me, but your browser may behave differently. Just don't test it on your 40 manga folder, use a folder with two or three to make sure it's working first.

 

[ediefalcopuppyalt]

I feel like the only real good use for this is blocking AI scrapers (which is good), but isn’t very effective when it comes to DDoS attacks since the attacker uses a whole network of bots of different ip addresses to send requests

 

[detarameMangadex]

Just curious: is this what has been making readers like Cubari.moe fail to load MangaDex chapters?

 

[AzrefReza]

I'm experiencing site issues that make me worry this enforcement goes beyond just excessive API calls; I've historically managed manga updates by organizing bookmark folders, since not everything's available all on one site, and once very couple weeks choosing the 'open all in tabs' option to pop fortyish followed series up at once and check through all of 'em for new chapters.

Now though, that seems to be triggering service timeout and hourslong blacklisting; they're all legitimate requests going through a consumer-standard browser, but because there's no apparent attempt to differentiate between 'this client made 40 requests in a second' and 'this client is making 40 requests in a second, every second, for a period indicative of scraping', suddenly junk's way less usable. Like, I'm already authenticating with a validated account and going through cloudflare half a dozen times a day too, how enshittified can the internet get???

E; I can't even access these forums or get a post through reliably - ctrl+F5 loads whatever page I'm targeting, but the moment I try to move to another page element it's back to 404ing me until I ctrl+f5 again - which obviously provides no avenue for using the post form, I had to log into and use an entirely different machine / browser since not even clearing cookies got me back in good graces.

Ha ha ha ha! I have the exact same problem for the exact same reason! From my experience, completely closing and reopening the browser (case in point, Firefox) makes the site usable again, but you can’t identify anymore (still blacklisted) and you can only open one tab at a time (IP still under heavy suspicion).

What I have done to avoid triggering the ban-hammer is reorganize my bookmarks in subfolders of roughly 15 links which I open one subfolder at a time. It also strongly mitigates the fact that Firefox pisses on itself and freezes my whole computer if I try to open 60+ tabs in one go.

But yeah, enshittification. It’s only a matter of time before I expand my “save-to-local-drive” app to work with MD and get rid of their bullshit…

 

[BestBoy]

What I have done to avoid triggering the ban-hammer

I'm pretty sure this is just regular rate limiting that's been in place for a while. Ask in the support channel on discord and you're likely to get a more complete answer from a mod.