Explain Anything

Home Page
Learning Path

Inspiration

Most “explanations” online come in one flavor: either overly simplified (and misleading) or so dense that you bounce off. I wanted a tool that can explain anything at multiple depths, but also make the explanation stick—not just sound smart for 30 seconds.

The idea for ExplainAnything came from a simple frustration: I could read a concept, feel like I understood it, and then fail to recall it a week later. So I built a tutor that doesn’t just explain—it helps you learn, practice, and revisit topics when your brain is about to forget.

What it does

ExplainAnything is an AI-powered learning app that turns any topic into a structured learning experience:

Multi-level explanations (beginner → intermediate → advanced) returned in consistent JSON so the UI can render it cleanly.
Creative learning modes (e.g., vivid story, rap, haikus, mnemonics, ELI5) that attack the same concept from different angles.
Study plan generator that creates a personalized day-by-day schedule with linked topics.
Spaced Repetition (SRS) so you can add topics you want to retain and review them based on when they’re due.

How we built it

Frontend (TypeScript + React): A web UI where users enter a topic, choose explanation/creative modes, generate study plans, and review due SRS cards.
API server: Routes that build high-quality prompts and request structured JSON outputs from an LLM (so the frontend can reliably display sections like misconceptions, examples, and step-by-step breakdowns).
Typed contracts: Generated API schemas/types (including Zod validation) so client/server stay in sync and UI rendering stays predictable.
Learning system: A local SRS engine that stores cards in the browser and updates intervals using a quality score (0–5), including due queues, forecasting, and stats.

Challenges we ran into

Getting reliable structure from AI output: “Creative” often means “messy.” The hardest part was enforcing strict JSON structure so the app stays stable.
Avoiding generic explanations: We had to push prompts toward concrete, non-template analogies and examples without sacrificing correctness.
Turning understanding into retention: Explaining is easy; helping users remember required adding SRS and making reviews feel lightweight instead of like homework.

Accomplishments that we're proud of

A single topic can be transformed into multiple learning artifacts (deep explanation + creative modes) while staying technically grounded.
The study plan + SRS loop turns one-off curiosity into an actual learning system.
Strong TypeScript typing across the app makes the experience feel cohesive and debuggable.

What we learned

Prompt design is product design: structure, constraints, and validation matter as much as model choice.
The best learning tools aren’t just informative—they’re repeatable and habit-forming.
Strong types and shared schemas reduce iteration time dramatically in a fast-moving hackathon environment.

What's next

Add quizzes/tests per topic and feed weak areas back into study plans automatically.
Account sync so SRS and progress persist across devices.
Better sharing/export (shareable “explain cards”, public study plans, and embeddable explain widgets).
Expand domains and add curated “learning paths” for common goals (interviews, CS fundamentals, system design).