Legal

AI Disclaimer

Which AI features Lio.gg offers, what is sent to Groq for them, and that we do not train models on that content.

Last updated: 12. Oktober 2026 · Lio.gg

1. Which features use AI

AI on Lio.gg is limited to features you start yourself or that a server operator turns on:

  • Server Builder: a Discord server draft from your wizard answers
  • Server Intelligence: an assessment from aggregated server facts, without chat history
  • Helpdesk triage: a suggestion for priority, category, macro, or help article on a ticket
  • Helpdesk AI support: a ticket reply based on the current message, recent turns, and published help articles

Other modules (moderation, logging, leveling, welcome, and the rest) do not use this language model. Lio.gg is not affiliated with Discord Inc.

2. What is sent

Server Builder sends your wizard answers: purpose, topics, style, size, selected modules, moderation options, and the server name if you provide one.

Server Intelligence sends aggregated facts: channel names and topics, module status, member and ticket counts, moderation stats, and role security signals. Member message history is not sent for this feature.

Helpdesk AI sends the ticket text (up to 1,500 characters), recent ticket messages in shortened form, and help articles, categories, and macros the server operator has published.

3. No training on your content

We do not use this content to train or fine-tune our own or third-party models. We do not sell it as training data, and we do not build training datasets from it.

The request is used only to produce the output you or the server operator asked for in that feature. Discord chats outside the helpdesk tickets described above are not used for AI.

4. Processor: Groq

Groq, Inc. (USA) processes these requests as a processor, solely to produce the output. A transfer to the USA may occur. How Groq handles API data is described in their privacy information: groq.com/privacy-policy. We do not control Groq, so we do not give our own guarantee about their internal processes.

The privacy description is in the Privacy Policy.

5. Outputs can be wrong

AI outputs can be incomplete, outdated, or wrong. They are suggestions, not a legal, security, or moderation decision. Server operators should review drafts, assessments, and ticket replies before relying on them or publishing them.

You remain responsible for text the bot posts because of your configuration. That is also stated in the Terms.

6. Contact

Questions about these features: support@lio.gg.

These texts are carefully drafted for the operation of Lio.gg and the Discord bot “Lio”, but do not replace individual legal advice. For questions contact support@lio.gg.