// Status
Every service below is checked live against the Discord API and the bot itself — no hard-coded 100%.
Running first check…
Every incident posted by the team, with start, end, duration and impacted service.
Every time the bot is updated, a new entry lands here.
The status page now polls the Discord API, the bot session and the guild directly instead of showing hard-coded uptime. Added a report button on every message plus a moderation queue with sanctions and role management.
Llama 3.3 70B is being sunset by Groq. The bot now runs on openai/gpt-oss-120b (served by Groq), with Qwen3 32B as a fallback. Prompts, commands and rate limits are unchanged — replies are noticeably sharper on reasoning-heavy questions.
Full English rewrite of the website. Added FAQ, Terms of Service, Privacy Policy, and this Status page. Copy-invite button with toast feedback.
Tightened the anti-bypass system prompt. Hard refusal of @everyone and @here pings is now also enforced server-side, not just in the prompt.
Replies now match the language of the +ask prompt automatically. Tested against fr, en, es, de, pt, it, nl, ja.
Per-user cooldown reduced from 6s to 4s on +ask. Server-wide soft cap kept at 30 requests/minute.
Windows scheduled from the admin dashboard where the bot or the site may be briefly affected.
Checks run at most once per minute and are cached between visitors to stay inside Discord rate limits.