Ticket AI is a premium, AI-native customer support platform: Support at the speed of intelligence. Automating ticket triage, skill-based routing, and response drafting using Google Gemini 2.0 Flash Lite.
Decoupled AI and email workflows using Inngest event-driven architecture, keeping Express response times at 20–50ms; optimized MongoDB queries via field projection and selective .populate() to prevent N+1 issues and reduce payload size.
Built a high-performance React 18 + Vite frontend featuring a Linear-inspired dark-mode landing page, skeleton loaders for zero layout shift, live AI triage state indicators, and role-based UI — architected for horizontal scaling via stateless JWT auth and MongoDB Atlas indexing.
- ✨ Linear-Inspired Landing Page: A professional, dark-themed first impression with smooth entrance animations and integrated feature mockups.
- 🧠 AI-Powered Triage & Analysis: Every incoming ticket is automatically analyzed by Gemini 2.0 Flash Lite for sentiment, technical requirements, and core issue identification.
- ⚡ Smart AI-Drafted Replies: Staff can generate high-quality, context-aware drafted responses with a single click, drastically reducing resolution time.
- 🛠️ Technical Skill-Based Routing: Tickets are automatically matched and assigned to moderators based on their specific technical expertise (React, Node, DB, etc.).
- 💬 GitHub-Style Resolution Thread: A professional, boxed message interface for clear and intuitive two-way communication between users and support staff.
- 📊 Real-Time Intelligence Dashboard: A glassmorphic admin interface featuring live ticket tracking, priority filtering, and message reply indicators.
- 🏗️ Reliable Event-Driven Architecture: Powered by Inngest for guaranteed background job execution, retries, and real-time observability.
- 🔒 Secure Role-Based Access: Granular permissions for Users, Moderators, and Administrators to manage workflows effectively.
- Frontend:
React 18,Vite,Tailwind CSS,Lucide Icons,Sonner Toasts,@vercel/analytics. - Backend:
Node.js,Express,Mongoose,JWT Authentication. - Cloud/AI:
Google Generative AI (Gemini 2.0 Flash Lite),Inngest(Event-driven background processes). - Database:
MongoDB Atlas(Optimized via indexing and lean queries). - Communication:
Nodemailer,Mailtrap.
- Node.js (v18+)
- MongoDB Atlas Account
- Google Gemini API Key
- Inngest Account (for local development)
# Clone the repo
git clone https://github.com/your-username/AI-ticket-Project.git
cd AI-ticket-Project
# Backend Setup
cd AI-ticket-Assistant
npm install
# Add .env: MONGO_URI, JWT_SECRET, GEMINI_API_KEY, ALLOWED_ORIGIN
# Frontend Setup
cd ../AI-ticket-Frontend
npm install
# Add .env: VITE_SERVER_URL- Inngest Dev:
npx inngest-cli@latest dev - Backend:
npm run dev(in Assistant directory) - Frontend:
npm run dev(in Frontend directory)
This project is designed to showcase how AI agents can handle complex, multi-step business workflows—from categorization to automated assignment and human-in-the-loop AI assistance.
Built with ❤️ by [Your Name]