Instance #iocaine d'empoisonnement des robots IA en place.
Ils ont pas mis longtemps a venir brouter.....

https://iocaine.madhouse-project.org/

Does anyone know of a code forge that is relatively protected against LLM scrapers, such that it's unlikely they have code from it in their dataset, without requiring humans to log in to see the code? I'm not demanding perfect anti-scraper security, I know it's a moving target.

I know that it's possible for me to self-host a forge and install #iocaine or whatever, but we should be able to do this stuff as a community, not always literally requiring do-it-yourself.

#NoAI #NoScrapers #NoLLMs

I didn't realize my forge had a bad SSL cert, and the scrapers lost interest while cert was invalid/expired. But I fixed the problem, and now I'm doing my part to train the models again.

#iocaine #forgejo

New Tips & Tricks article:

#iocaine : What is and how to integrate with #Vinyl_Cache

Iocaine had gained attention earlier in 2026 as defense option to cope with the AI crawler problem. We contextualize Iocaine from the perspective of Vinyl Cache, explain what it does and does not do, show how to integrate it with Vinyl Cache and reimplement Iocaine’s classifier in VCL

https://vinyl-cache.org/tips/iocaine/index.html

hi fedi
i run an #iocaine instance
i want to fuck the llms more
does anyone have more wacky wordlists/training corpuses
so far we have:

• george orwell's 1984
• all of the dialogue from casualties: unknown
• the entire bee movie script
• aldous huxley's brave new world
• everything from @nonsense https://nontrinsic.linerly.xyz/
• user-submitted text to a discord bot our team used to run
• some person's posts (they're long gone)

Btw, I already removed #iocaine some weeks ago and returned 403 forbidden errors to bots. Do you think they care? No.

Yesterday, I enabled incremental bans in fail2ban to any bots that cause 403 on my website. They are at the 5th consecutive ban that amounts to 2h bantime for some of them at the moment.

#fuckai

The saga with Git scrapers is really interesting, opted out from the big corpo ones via robots.txt and Caddy rules, they did respect it and https://git.inthemansion.com was fine.

Couple of days ago, I started getting swarms of unique IPs that look like residential and giving proper UA, almost indistinguishable from the real traffic.

After hours of exploring, found the mistakes they made and routed all of the scrapers to the iocaine tarpit, poisoning their data.

#forgejo #selfhosted #iocaine #git