Yo. You're in my corner of the internet.
I go by Severian in a lot of places like Discord and Reddit. I'm trying to get into image gen and model finetuning, as well as bot making. Site's not halfway done yet, too.
If you're here though, you're probably after the various shit I've put up online or information I curate, mostly AI/RP-related. Check them out below:
- GitHub
- github.com/severian-dev
- Hugging Face
- huggingface.co/finis-est
Contact
My email is yo@[this domain] should you want to contact me. Hope you find what you're looking for.
漏 2026 - now get the hell out of here 馃憠
Lately
Activity
Two successors, one showcase
GPT-6.2-Astra and GPT-6.2-Sol are taking over from GPT-6-Astra and GPT-6.1-Sol in my setup. I gave them the three showcase rounds the others ran, word for word, and put each piece beside the one it follows.
by Jonas
- Transcript The showcase prompt, round 3, GPT-6.2-Astra
- Transcript The showcase prompt, round 3, GPT-6.2-Sol
Five models, one mech and one character
Seven models, one showcase
Fable 5.1, Fable 5.2 and Opus 5.5, one idea in the queue
Fable 5.1 and 5.2, one design review
Three agents, one face
Orgs

trashpanda-org
a finetuning organization

Avalon
a botmaker collective

exvanitas
my personal workshop
Projects
the swarm, as it happens
what the swarm has learned, kept
Char card extractor for you-know-what
mirrors trashpanda.land hitani.me portalnexus.link lemuria.dev ladyofpa.in
Repos
SillyTavern extension: reasoning prefill + preserved thinking for Kimi/Moonshot via API partial mode
SillyTavern extension: Fish Audio voices in the built-in TTS, with a voice browser and free sample previews.
SillyTavern server plugin: Fish Audio API bridge for the Fish Audio TTS extension. Holds the key, streams audio.
various notebooks, idk