preqly

Inspiration

Students often fall behind because courses feel linear when they’re actually built on dependencies. We wanted to make those hidden topic relationships visible.

What it does

Preqly turns a static syllabus into a dynamic, AI‑powered learning ecosystem. Instead of just visualizing prerequisites, it uses a coordinated team of specialized agents that teach, reinforce, and evaluate every concept in the course. Students don’t just see what they need to learn — they actively move through it with guided support.

🧭 Interactive prerequisite map

Preqly converts any syllabus into a visual graph that reveals hidden topic dependencies. Students can:

Explore how concepts connect
See what each topic unlocks next
Track progress across the entire course
Personalize their map and workspace

Every node in the graph becomes an entry point into an actual learning experience.

🤖 Multi‑agent learning system

Preqly is built around three collaborative AI agents, each focused on a different stage of learning.

🎓 Teach Agent — Learn the topic

The Teach Agent turns each concept into a focused, digestible lesson.

Generates short teaching videos for each topic
Adapts explanations to the student’s current knowledge and prerequisites
Emphasizes how the concept connects to earlier topics and what it unlocks next
Makes every node in the map feel like a self‑contained learning module

🧠 Flashcard Agent — Remember the topic

Once a concept is introduced, the Flashcard Agent reinforces it through active recall.

Automatically generates flashcards for definitions, examples, and key relationships
Supports spaced‑repetition‑style review flows
Adjusts card difficulty based on how well the student is doing
Helps students build durable, long‑term memory of each concept

📝 Test Agent — Master the topic

To verify understanding, the Test Agent creates targeted assessments around the map.

Generates quizzes and practice questions for individual topics or clusters of related concepts
Varies difficulty from basic checks to exam‑style problems
Surfaces weak areas directly on the prerequisite map
Uses results to suggest what to review next with the Teach and Flashcard agents

Together, these three agents turn a static prerequisite graph into a closed learning loop: learn → reinforce → test → adapt, all grounded in the structure of the original syllabus.

How we built it

We built Preqly with Next.js, React, and Supabase. The frontend renders an interactive graph workspace, while Supabase handles auth, storage, persistence, and background course-generation jobs from uploaded PDFs.

Challenges we ran into

The biggest challenge was making the PDF-to-course pipeline reliable. We had to debug stuck background jobs, function auth issues, and make the system resilient when uploads or processing failed.

Accomplishments that we're proud of

We’re proud that Preqly is more than a static visualization. It supports interactive graph editing, saved layouts, progress tracking, version history, and a full course workspace around the map.

What we learned

We learned a lot about designing around async systems: background jobs, error states, retries, and keeping the UI honest when processing is still happening behind the scenes.

What's next for Preqly

Next, we want to improve syllabus parsing accuracy, add smarter topic extraction, expand collaborative features, and make the workspace more useful across multiple classes and study workflows.