p-e-w/heretic — Fully Automatic Censorship Removal for LLMs (14.3K Stars, +1,066/day)

GitHub·high signal

heretic is a Python tool that automatically removes safety filtering from language models, achieving full uncensoring without manual jailbreaks. At 1,066 stars/day and 14,328 total it is one of the fastest-growing repos on GitHub today, signaling sustained demand for unrestricted local model execution. Critical dual-use signal: builders running local models for security research, red-teaming, and autonomous pipelines will find this relevant; it also raises the agent-security threat surface.