☁️ AI Cloud Cost Optimiser

An AI-assisted Azure cost optimisation platform built with React, FastAPI, SQLite, Azure CLI, and Google Gemini. It scans Azure resource groups, identifies potential cost-saving opportunities, and presents actionable recommendations with estimated savings.

---

📑 Table of Contents

• Overview
• Architecture
• Tech Stack
• Project Structure
• Quick Start
• Azure Configuration
• AI Analysis
• API
• Database
• Troubleshooting
• Limitations
• License

---

🧭 Overview

AI Cloud Cost Optimiser helps users analyse Azure resource groups and find potential cloud cost-optimisation opportunities.

Users can:

• Create an account and log in
• View available Azure resource groups
• Select a resource group for analysis
• Scan Azure resources
• Analyse resources with Google Gemini
• View optimisation recommendations and estimated savings
• Review previous analyses

The application provides recommendations for review and does not automatically execute Azure changes.

---

🏗️ Architecture

┌────────────────────────────────────────────────────────────┐
│                 AI CLOUD COST OPTIMISER                    │
├────────────────────────┬───────────────────────────────────┤
│       Frontend         │             Backend               │
│                        │                                   │
│ React + TypeScript     │ FastAPI                           │
│ Vite                   │ Authentication                    │
│ React Router           │ Azure Scanner                     │
│ Axios                  │ Gemini AI Analyzer                │
│ Tailwind CSS           │ SQLite                            │
└────────────────────────┴───────────────────────────────────┘
                              │
              ┌───────────────┴───────────────┐
              ▼                               ▼
       Microsoft Azure                 Google Gemini
          Azure CLI                    AI Analysis

Analysis Flow

Select Resource Group
        ↓
Azure Resource Scan
        ↓
Gemini Analysis
        ↓
Save Analysis
        ↓
Optimisation Report

---

🛠️ Tech Stack

Layer	Technology
Frontend	React, TypeScript, Vite
UI	Tailwind CSS
Backend	Python, FastAPI, Uvicorn
Authentication	JWT, bcrypt
Database	SQLite, aiosqlite
Cloud	Microsoft Azure, Azure CLI
AI	Google Gemini (google-genai)
Real-time	WebSockets
HTTP Client	Axios

---

📁 Project Structure

AI_Cloud_Cost_Optimiser/
│
├── backend/
│   ├── main.py
│   ├── auth.py
│   ├── db.py
│   ├── azure_scanner.py
│   ├── ai_analyzer.py
│   ├── requirements.txt
│   └── .env.example
│
├── frontend/
│   ├── src/
│   │   ├── App.tsx
│   │   ├── main.tsx
│   │   ├── index.css
│   │   ├── components/
│   │   │   ├── Navbar.tsx
│   │   │   └── ProgressTracker.tsx
│   │   └── pages/
│   │       ├── Login.tsx
│   │       ├── Signup.tsx
│   │       ├── Dashboard.tsx
│   │       ├── History.tsx
│   │       └── Report.tsx
│   ├── package.json
│   └── vite.config.ts
│
├── .gitignore
└── README.md

---

🚀 Quick Start

Prerequisites

• Python 3.11+
• Node.js 18+
• npm
• Azure CLI
• Azure account with access to the resources
• Google Gemini API key

Check the installations:

python --version
node --version
npm --version
az --version

Backend

cd backend

python -m venv venv
venv\Scripts\activate

pip install -r requirements.txt

Create backend/.env from .env.example:

GEMINI_API_KEY=your_gemini_api_key
GEMINI_MODEL=gemini-2.5-flash
JWT_SECRET=your_secret

Start the backend:

uvicorn main:app --reload

API:

http://localhost:8000

Swagger documentation:

http://localhost:8000/docs

Frontend

Open a second terminal:

cd frontend
npm install
npm run dev

Open:

http://localhost:5173

---

☁️ Azure Configuration

Authenticate with Azure:

az login

Check the active account:

az account show

List available subscriptions:

az account list -o table

Select a subscription if required:

az account set --subscription "<SUBSCRIPTION_NAME_OR_ID>"

Verify resource groups:

az group list -o table

The current scanner uses the Windows Azure CLI path:

C:\Program Files\Microsoft SDKs\Azure\CLI2\wbin\az.cmd

If Azure CLI is installed elsewhere, update the path in:

backend/azure_scanner.py

---

🤖 AI Analysis

Google Gemini analyses the Azure resource information collected by the scanner.

The application can identify opportunities involving areas such as:

• Virtual machine configuration
• Idle or unnecessary resources
• Auto-shutdown
• Spot allocation
• Managed disks
• Unattached disks
• Storage redundancy
• Storage access tiers
• Lifecycle policies
• Other configuration-based cost opportunities

The generated report includes recommendations and estimated savings.

Estimated savings are not guaranteed billing savings and should be validated before making infrastructure changes.

---

🔌 API

Method	Endpoint	Purpose
POST	/api/auth/signup	Create account
POST	/api/auth/login	Login
GET	/api/resource-groups	Get Azure resource groups
POST	/api/analyze/start	Start an analysis
POST	/api/analyze	Run analysis
GET	/api/history	Get analysis history
WS	/ws/progress/{analysis_id}	Analysis progress

Full API documentation is available at:

http://localhost:8000/docs

---

🗄️ Database

The application uses SQLite with aiosqlite.

The database is created locally as:

backend/app.db

It stores user accounts and completed analysis information, including resource group, status, findings, estimated savings, and analysis results.

---

🐛 Troubleshooting

Azure CLI not found

az --version

If Azure CLI works in the terminal but not in the application, check the path in backend/azure_scanner.py.

Azure authentication problems

az login
az account show

Make sure the selected Azure subscription contains the resource group you want to analyse.

Gemini API key error

Check:

backend/.env

and verify:

GEMINI_API_KEY=your_gemini_api_key

Restart the backend after changing the .env file.

Frontend cannot connect to backend

Make sure the backend is running:

uvicorn main:app --reload

Then check:

http://localhost:8000/docs

---

⚠️ Limitations

• Savings are estimates, not guaranteed Azure billing savings.
• The application focuses on configuration-based optimisation rather than complete Azure Cost Management analysis.
• Comprehensive long-term CPU/memory utilisation analysis is not currently provided.
• Azure CLI recommendations are not automatically executed.
• The current Azure CLI path is Windows-specific.
• SQLite is intended for the current local/development use case.

---