SafeChat AI
Chat with Mia - safe, rate-limited, anti-hallucination, and opt-out friendly.
SafeChat AI
SafeChat is how you talk to Mia. It runs her conversational replies through a large language model, with safety, rate limits, and an anti-hallucination knowledge lookup so answers stay grounded.
How to Chat
Talk to Mia in a whitelisted channel or in DMs (where enabled). She keeps a short conversation history for context, so replies follow the thread.
What Keeps It Safe
- Cooldowns and hourly limits prevent spam.
- Knowledge lookup pulls factual context from a knowledge base, so Mia is less likely to make things up.
- Sentiment-aware replies, and a sassier side when the moment calls for it.
- Mia has a mood that colors her tone - see Mood.
Your Privacy
- Mia maintains conversation history to give good replies.
- You can opt out of having your conversation data stored. Once you opt out, Mia respects it.
- Ask staff through Support if you want help opting out or clearing data.
Common Actions
Check Mia's mood
%mood
Opt out of data storage
Mia provides an opt-out system that respects your privacy. Reach out via [Support](/support) if you need a hand.Troubleshooting
| Symptom | Likely Cause | Next Step |
|---|---|---|
| Mia is not replying | Cooldown, hourly limit, or a non-whitelisted channel | Wait a bit, or try a whitelisted channel |
| Mia will not chat in DMs | DM replies may be disabled | Use a server channel |
| Mia got a fact wrong | LLMs can still slip | Let staff know; knowledge lookup improves over time |