Adhyan

Inspiration

Studying from large chunks of text, dense PDFs, or long YouTube lectures often feels overwhelming and inefficient. We wanted to build a platform that transforms raw learning material into structured, interactive, and engaging study resources. The idea was to combine AI with modern web technologies to reduce the effort of content digestion and help learners focus on understanding, recalling, and applying concepts.

What it does

Adhyan is an AI-powered study platform that takes your PDFs, text, or YouTube videos and turns them into smart study materials. It automatically generates:

Structured Notes with a hierarchical table of contents
Flashcards (Definitions, Recall, and Application types)
Interactive Quizzes with instant feedback
Mind Maps for visualizing concept relationships

The platform ensures that content is not just summarized, but restructured into formats that actively support better learning and retention.

How we built it

Frontend: Next.js, React, TypeScript, Tailwind CSS for a smooth, modern user interface.
Backend: Next.js API Routes with Prisma ORM and PostgreSQL for structured data storage.
AI Integration: Google Gemini 2.0 Flash connected via LangChain, with carefully engineered prompts for different study materials.
Authentication: NextAuth.js for secure and seamless user login.
Content Pipeline:
1. Users upload PDFs, input text, or provide YouTube links.
2. System extracts and preprocesses raw text.
3. Content is chunked, scored, and filtered for quality.
4. AI generates notes, flashcards, quizzes, and mind maps.
5. Materials are stored and displayed in an interactive dashboard.

Challenges we ran into

Designing a robust content preprocessing pipeline that could handle PDFs, text, and YouTube transcripts reliably.
Extracting Youtube Videos Transcripts was difficult as there are no free API options
Engineering context-rich prompts to get consistent and high-quality outputs from Gemini.
Structuring flashcards, quizzes, and mind maps in a way that feels natural and educationally effective.
Balancing AI generation cost and speed while ensuring good performance at scale.

Accomplishments that we're proud of

Built a multi-format study platform that integrates text, video, and documents into one streamlined workflow.
Developed a content scoring system that prioritizes high-value material for quizzes and flashcards.
Successfully implemented interactive mind maps that visualize relationships between concepts.

What we learned

How to effectively design and fine-tune prompt engineering strategies for different output types.
The importance of preprocessing and filtering before feeding data to AI to avoid irrelevant or low-quality results.
How to structure a full-stack AI project, balancing UX, backend robustness, and AI model integration.
Learned that learners value active recall tools (quizzes, flashcards) much more than just passive summarization.

What's next for Adhyan

Adding collaborative study packs, where groups can share notes, flashcards, and quizzes.
Expanding support for more content formats (Word docs, lecture slides).
Enhancing quizzes with adaptive difficulty based on learner performance.
Building a mobile app for on-the-go studying.
Exploring multilingual support to make Adhyan more accessible worldwide.