JEMPOL - Integrated Information Chatbot RAG
JEMPOL (Jaringan Edukasi dan Informasi Layanan Mandiri POLIJE) is an intelligent chatbot I built for students and staff at Politeknik Negeri Jember. Think of it as your always-available information desk—ask about academic programs, student services, or anything else about the campus, and JEMPOL gives you accurate answers instantly, without the back-and-forth with administration.
What JEMPOL Does
- Knows everything about your campus – Questions about programs, policies, deadlines, services? JEMPOL has answers.
- Learns from your documents – Uses RAG (Retrieval Augmented Generation) to pull from official resources, so answers are always accurate.
- Understands what you’re asking – Semantic search with cosine similarity means it gets your question even if you don’t use perfect wording.
- Blazingly fast – Built on Pinecone for instant responses without waiting.
- Feels natural – Clean interface, real-time conversation, no confusing forms to fill out.
Screenshots
Built With
Frontend & Framework
- Next.js – React framework that keeps things fast and organized
- Shadcn UI – Clean, accessible components that work everywhere
- Tailwind CSS – Styling that doesn’t bloat the code
- Framer Motion – Smooth animations that don’t feel sluggish
- Lucide React – Beautiful icons throughout the interface
AI & Data
- RAG (Retrieval Augmented Generation) – Pulls information from knowledge base for accurate answers
- Semantic Search – Uses cosine similarity to understand what you really mean
- Pinecone – Vector database that makes searching instant
- Google Generative AI – Powers the conversational AI
- LangChain – Orchestrates the AI workflows
Utilities
- React Markdown – Renders formatted content cleanly
- Zod – Schema validation to catch errors early
Try It Out
Head to jempol.side.my.id to see it in action. Just ask JEMPOL anything about campus—it’s ready to help.
Contribution
This is a personal project and does not accept external contributions.