AI Reading Companion

An AI-powered reading comprehension tool that guides learners through a passage with personalised questions and written feedback. The flow is: 1. Student picks a difficulty level (Beginner / Intermediate / Advanced) 2. A passage about honeybee biology is split into 4 short sections 3. After each section, 2 AI-generated comprehension questions appear 4. The student types a free-text answer. The AI scores it 0–100 with written feedback 5. Wrong answers surface the relevant excerpt from the passage alongside the model answer 6. After the final section, a synthesis question asks the student to integrate everything they have read 7. A results page shows overall score, time taken, every answer with the model answer, and a "level up" option if they scored ≥ 85% The app is built on Next.js 15 App Router with React 19 and TypeScript in strict mode. All AI interaction goes through a server-only module wrapping the Anthropic SDK. On the first request for a given difficulty, a single API call generates all 8 comprehension questions in one prompt — the result is cached for the process lifetime, making sections 2–4 instant cache hits. Session data is serialised to sessionStorage before routing to the results page, bridging React context across the route transition without requiring a database.