v2.3.3 just shipped and it's probably the biggest update so far.
use Qwen 3.6 now with remote Access, codex and plugins such as cavemen!
Locally Uncensored
our users casually generate uncensored images and videos on a 6GB GPU while billion dollar companies still can't let you say 'draw a person holding a knife' without a content warning
v2.3.0 is out.
comfyui plug & play — auto-detects, one-click install, zero config. image-to-video on 6gb vram. uncensored image gen. glm 5.1 + gemma 4 support.
one app. chat, code agent, image gen, video gen. no cloud, no docker.
open source (agpl-3.0)
https://t.co/95QyU2L5K7
we shipped web search for local AI agents yesterday because someone asked for it in a github discussion
the whole feature request → shipped pipeline was about 12 hours
Locally Uncensored
Locally Uncensored v2.1 — now with Gemma 4 support on day one.
The Gemma 4 Benchmarks jumped from 6% to 86% on agentic tasks.
Download and run — that's it.
I literally have to twist people’s arms to try Claude Code.
It’s the strangest thing: here’s basically a genie in a bottle that can do anything for you.
Yeah yeah I’ll try it tomorrow.
@etiennebuilds @EHuanglu Sadly no such thing as a digital twin generated by a single LLM. Atleast as of right now. And its not even close for complex plans..