Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{200284,
        author = {Saikarthik M and Sharath S and Augustine Fletcher A. S},
        title = {Smart Education System: A Full-Stack E-Learning Platform with Embedded Generative AI, Multimodal Content, and Learning Analytics},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {2056-2061},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=200284},
        abstract = {We started this project after noticing something annoying — every e-learning platform we used kept the study material and the AI help completely separate. You read a PDF, you get stuck, you open ChatGPT in another tab, paste the content, get an answer, come back. It works, but it is clunky and the instructor has zero visibility into any of it.
So, we built one platform that keeps everything together. Students can browse subject materials, watch videos, and use AI tools like summarization, quiz generation, and diagram creation without ever leaving the page. Teachers get an upload workflow. Everyone gets a dashboard with XP and activity tracking.
We used Next.js for the full stack, MongoDB for data, and Cloudinary for media storage. For AI we picked Google Gemini — mainly because it has a free API, which mattered for a student project with no budget. All API calls happen server-side so keys never reach the browser.
Building it was not always smooth. Gemini’s JSON mode occasionally returned output wrapped in markdown blocks instead of raw JSON, which broke our quiz renderer until we added a sanitization step. Cloudinary uploads failed silently when we ran them in parallel with database writes. JWT tokens were inconsistent across routes for a while and caused authenticated users to get 401 errors on the dashboard — that one took us almost two days to track down.
This paper covers the architecture, how we handled these problems, what the system looks like now, and what we would do differently.},
        keywords = {Smart education, e-learning, generative AI, Next.js, learning analytics, multimodal content, full-stack web application.},
        month = {May},
        }