PadhAI (ExamSaathi)

PadhAI (ExamSaathi) is an AI-powered study companion built specifically for Indian students preparing for board exams, competitive tests like JEE/NEET, or university courses. It removes study friction by letting students point their phone camera at a dense textbook page or handwritten notes, select their preferred language (Hindi or English), and instantly receive a plain-language explanation, tailored practice questions with answers, and a concrete revision tip.

Built for the First Commit hackathon, the project utilizes a serverless architecture scaling completely to zero when idle, keeping costs minimal while delivering fast, multimodal AI tutoring.

Core Features & Workflow

1. Multimodal Upload: Snap a photo of a textbook page, upload handwritten notes, or paste text directly into the clean, minimalist mobile-friendly interface.

2. Bilingual Support: Tailored explanations available in simple Hindi + English mix or standard English.

3. Actionable Insights: Instantly generates a structured explanation, active practice questions to test understanding, and a clear next step for revision.

4. Save to Weak Topics: Easily bookmark tricky concepts to review later on the dedicated History page.

Technical Architecture

• Frontend: Built using Next.js 15 App Router, TypeScript, and Tailwind CSS, featuring a focused three-screen flow (Home/Upload, Result, and History) deployed on AWS Amplify Hosting.

• Backend: Powered by three single-purpose Python 3.12 AWS Lambda functions behind an Amazon API Gateway HTTP API:

  • POST /analyze: Validates uploads, handles temporary image storage, and interacts with Amazon Bedrock.

  • POST /history & GET /history: Saves and retrieves study sessions via Amazon DynamoDB using secure session identification.

• AI Engine: Powered by Amazon Bedrock using Claude (via system-defined inference profiles) to seamlessly handle both text and vision inputs through the Converse API.

• Security & Storage: Employs temporary private Amazon S3 storage with automatic lifecycle expiry, along with anonymous session management via Amazon Cognito Identity Pools.