專案旨在為校園通識實作競賽活動「校原尋禮」開發一款名為「AI 小精靈」的活動對話助理。系統架構結合了 Django 網頁框架、OpenAI LLM API、Supabase (PostgreSQL) 資料庫以及 Chroma 向量資料庫，並部署於 Google Cloud Run 雲端平台。

透過 RAG（檢索增強生成）技術與 Tool Function 功能，「AI 小精靈」能夠讓參與活動的玩家以自然語言查詢個人的遊戲關卡進度、積分、排行榜名次，或是檢索活動規則與公告，從而提升玩家在校園闖關活動中的科技素養與互動體驗。 

This project aims to develop an AI conversational assistant named “AI Sprite” for the campus general education competition event Campus Quest. The system architecture integrates the Django web framework, the OpenAI LLM API, a Supabase (PostgreSQL) database, and a Chroma vector database, and is deployed on Google Cloud Run.

By leveraging Retrieval-Augmented Generation (RAG) techniques and function calling capabilities, “AI Sprite” enables participants to naturally query their game progress, accumulated points, and leaderboard rankings, as well as retrieve event rules and announcements. This enhances participants’ technological literacy and interactive experience within the campus scavenger hunt activity.