
AI question answering · RAG-powered · Self-hosted · Zero server lag · AI moderation
Your players have thousands of questions. You can't answer them all.
New players join your server and immediately ask the same questions.
Where do I start? How does this game mode work? What are the rules?
Until you have staff and a community, those questions go unanswered.
Players leave before they understand what makes your server worth staying for
Battista stays online 24/7. It knows your server because you teach it.

Architectural overview and functional capabilities of the module.
MOD.04-a
Answers are based entirely on your server documentation. Fast. Accurate. Context-aware.
MOD.04-b
Battista monitors chat continuously. Toxic messages are detected and removed automatically. Keeps your community clean before you have moderators to do it manually.
MOD.04-c
Upload custom guides, rules, FAQs, and lore. Battista uses Retrieval-Augmented Generation to ground every answer in your actual content. Not generic Minecraft knowledge.
MOD.04-d
All AI processing runs on the separate backend. The plugin is asynchronous. Your TPS stays clean.
FEED IT YOUR DOCS
Upload your guides, rules, FAQs, and game mode documentation. Battista indexes everything and builds its knowledge base from your content. It only answers what you've taught it. No hallucinations about your server.
DEPLOY TO YOUR SERVER
Install the Spigot plugin on your server. Connect it to the self-hosted backend via config.yml. Done. Players can start asking immediately.
IT RUNS ITSELF
Players ask via command, chat tag, or natural question detection. Battista responds in real time with zero server lag. It also monitors chat for toxic content and removes it automatically.
Both components are required. Setup takes under 30 minutes with Docker.
No. Battista only answers based on documents you provide. If it doesn't know, it says so.
Under 30 minutes with Docker. The repository includes step-by-step instructions.
Configurable. Supports multiple providers with model rotation and fallback. Local models are supported if you want zero API costs.
No. All AI processing happens on the backend. The plugin uses async calls. Your server TPS is unaffected.
Yes. Both components are open source and MIT licensed. You cover your own hosting and LLM API costs.