CypherCypher 5.0 - AI-Powered Ticket System

A support ticket system with an AI agent that automatically analyzes tickets and suggests solutions.

---

How It Works (Simple Version)

┌─────────────┐      ┌─────────────┐      ┌─────────────┐
│   User      │ ──▶  │   Backend   │ ──▶  │   AI Agent  │
│  (Browser)  │ ◀──  │  (FastAPI)  │ ◀──  │  (Gemini)   │
└─────────────┘      └─────────────┘      └─────────────┘

1. Users Submit Tickets

• A user visits the website and logs in
• They fill out a form describing their problem
• The ticket gets saved to the system

2. Support Team Views Tickets

• Dev/support users log in and see all tickets
• They can click "Ask AI" on any ticket
• This opens the Agent Analysis Panel

3. AI Agent Analyzes the Ticket

The AI agent does 5 things automatically:

Step	What It Does
Read	Reads the ticket content
Extract	Pulls out key info (merchant ID, error codes)
Search Logs	Looks up related system logs
Find Docs	Searches documentation for relevant guides
Diagnose	Uses Google Gemini to create a diagnosis and suggested reply

4. Support Gets a Ready-to-Send Reply

• The AI shows its thinking process step-by-step
• Provides a diagnosis with confidence score
• Suggests a reply that can be edited and sent

---

Tech Stack

Part	Technology
Frontend	Next.js, React, Tailwind CSS
Backend	Python, FastAPI
AI Agent	LangGraph + Google Gemini
Database	Supabase (users), In-memory (tickets)

---

Quick Start

Backend

cd backend
pip install -r requirements.txt
python -m uvicorn main:app --reload

Runs on http://localhost:8000

Frontend

cd frontend
npm install
npm run dev

Runs on http://localhost:3000

---

Supabase Setup

1. Create a Supabase project at https://supabase.com
2. Run this SQL in the Supabase SQL Editor:

CREATE TABLE cc_users (
  id UUID DEFAULT gen_random_uuid() PRIMARY KEY,
  name TEXT NOT NULL,
  email TEXT UNIQUE NOT NULL,
  password TEXT NOT NULL,
  role TEXT NOT NULL DEFAULT 'user' CHECK (role IN ('user', 'dev')),
  created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW()
);

-- Add sample users
INSERT INTO cc_users (name, email, password, role) VALUES
  ('Arya', 'arya@gmail.com', 'password123', 'user'),
  ('Arhaan', 'arhaan@gmail.com', 'devpass456', 'dev');

---

Environment Variables

Create backend/.env:

SUPABASE_URL=your_supabase_url
SUPABASE_KEY=your_supabase_key
GOOGLE_API_KEY=your_gemini_api_key

---

Project Structure

cyphercypher5.0/
├── backend/
│   ├── main.py          # FastAPI app (login, tickets)
│   ├── agent.py         # LangGraph AI agent
│   ├── router.py        # /agent/analyze endpoint
│   ├── mock_db.py       # Fake logs & docs for testing
│   └── requirements.txt
│
├── frontend/
│   ├── app/
│   │   ├── page.js      # Main page (login, tickets)
│   │   └── globals.css
│   └── components/
│       └── AgentInsightPanel.tsx  # AI analysis modal

---

How the AI Agent Works (Technical)

The agent uses LangGraph to run a sequence of steps:

extract_metadata → search_logs → search_docs → generate_solution

Each step updates a shared state object:

{
  "ticket_text": "...",
  "merchant_id": "m_123",
  "logs_found": ["Error: 403..."],
  "relevant_docs": ["API Guide..."],
  "diagnosis": "The API key is invalid...",
  "confidence_score": 0.85,
  "recommended_action": "Dear customer...",
  "steps_log": ["Found merchant ID", "Checked logs", ...]
}

The final generate_solution step calls Google Gemini with all the gathered context to produce a diagnosis and suggested reply.

---

License

MIT