AI Study Buddy: A Full-Stack Intelligent Learning Assistant Based on FastAPI and React

AI Study Buddy is an open-source full-stack intelligent learning assistance platform that combines a FastAPI backend and React frontend. It uses GPT-4o-mini to generate notes, quizzes, and flashcards, and supports PDF export, JWT authentication, and machine learning-based topic classification. The project is maintained by Shubh50575, with source code hosted on GitHub (link: https://github.com/Shubh50575/AI-Study-Buddy), and was released on June 15, 2026.

With the popularization of digital learning, students face issues of information overload and low efficiency. Traditional learning methods struggle to meet personalized needs, while commercial software is often single-functioned and expensive. AI Study Buddy emerged as an open-source full-stack intelligent assistant that combines large language models with traditional learning tools to provide personalized, interactive learning experiences, forming a complete learning assistance workflow from note generation to topic classification.

The backend is based on Python 3.13+ and the FastAPI framework, leveraging asynchronous features to handle concurrency efficiently. SQLite + SQLAlchemy is used for lightweight data persistence. AI capabilities are integrated via the OpenRouter API to access GPT-4o-mini, with a unified interface facilitating model switching and reducing security risks. The frontend uses React 18 + Vite, with a component-based architecture to enhance reusability, and LocalStorage to store JWT tokens to adapt to learning scenario needs.

1. Intelligent Content Generation: Input a topic to call GPT-4o-mini for generating notes, quizzes, and flashcards, with multi-modal output adapting to different learning styles; 2. Keyword Extraction and Topic Classification: RAKE algorithm for keyword extraction, TF-IDF + Naive Bayes for topic classification; 3. Export and Sharing: Supports PDF/TXT format export, breaking the closed nature of learning content.

JWT authentication is used with bcrypt password hash storage, forming a complete security loop including registration, hashing, token generation, storage, and verification. JWT tokens are stored in LocalStorage, which has XSS risks, but these are mitigated through content security policies and input validation. In production environments, tokens can be migrated to httpOnly Cookies.

Applicable scenarios include self-learners (quickly generating introductory materials), exam-preparing students (self-assessment), teachers (lesson preparation assistance), etc. The open-source nature supports secondary development, allowing expansion of more export formats, AI models, or collaboration features. It is both a ready-to-use tool and an evolving platform.

AI Study Buddy combines AI with traditional learning tools to enhance learning experiences rather than replace traditional methods. In the future, we can expect more intelligent context understanding, precise learning path recommendations, and natural interactions. The project demonstrates AI integration practices for developers, provides efficiency tools for learners, and proves the potential of combining AI with education.