CodeCoach AI

An AI-powered code review platform that helps beginner programmers improve their Python code through automated feedback, scoring, and personalized explanations.

Live Demo: https://codecoachai.onrender.com

---

Overview

CodeCoach AI analyzes Python programs using a combination of:

• Rule-based static analysis
• AI-generated explanations (powered locally by Ollama)
• Code quality scoring
• Personalized learning suggestions

The goal is to help students understand why their code can be improved—not just whether it works.

The deployed version demonstrates the application's interface and rule-based analysis. The full AI tutoring experience is available when running the project locally with Ollama.

---

Screenshots

Home

---

Analyze Code

---

📊 Results — Overview

📊 Results — Code Comparison

---

History

Features

• AI-generated code explanations
• Rule-based bug detection
• Code quality score (0–100)
• Personalized learning tips
• Side-by-side code comparison
• Submission history
• User authentication
• Public web deployment with Render

---

Tech Stack

Backend

• Python
• Flask
• SQLite

AI

• Ollama
• Llama 3.2

Frontend

• HTML
• CSS
• Jinja2

Tools

• Git
• GitHub
• Render
• Pytest

---

Project Structure

CodeCoachAI/
│
├── app.py
├── requirements.txt
├── README.md
│
├── services/
│   ├── analyzer.py
│   └── ai_service.py
│
├── templates/
│
├── static/
│
├── tests/
│
└── screenshots/

---

Running Locally

Clone the repository:

git clone https://github.com/Raniafsl/CodeCoachAI.git
cd CodeCoachAI

Create a virtual environment:

python3 -m venv .venv
source .venv/bin/activate

Install dependencies:

pip install -r requirements.txt

Install and start Ollama.

Pull the model:

ollama pull llama3.2:3b

Start Ollama:

ollama serve

Run the application:

python app.py

Open:

http://127.0.0.1:5000

---

Testing

Run the unit tests:

pytest

---

Deployment

The application is deployed using Render.

The public deployment demonstrates the user interface and rule-based analysis.

The complete AI tutoring functionality is available when running the application locally with Ollama.

---

Future Improvements

• Support multiple programming languages
• Richer static code analysis
• PostgreSQL database
• Docker support
• Cloud-hosted AI backend
• Instructor dashboard
• Syntax highlighting in the editor

---

Author

Rania Faisal

Computer Science Student,
University of Waterloo

GitHub: https://github.com/Raniafsl