{% extends "base.html" %} {% block nav_settings %}active{% endblock %} {% block title %}Settings{% endblock %} {% block content %}
Configure the LLM backend, email (SMTP), proxy, and resume.
Default: nightmare Ollama (qwen3.8fast). The pipeline falls back to keyword matching if the model is saturated.
Runs every 6h via systemd timer (procyon-autopilot.timer). When Auto-send is On, prepared apps are emailed to their scraped contact automatically (audit + fabrication guards still apply). Jobs with a URL but no email still need you to apply via the posting.