link: https://youtu.be/gf7XsH3yofg

Inspiration

After taking physics classes, we realized that solving homework was often the hardest part of learning. Most AI tools and homework apps simply gave us the answer or walked through the entire solution, making it easy to copy without actually understanding the concepts. What helped us the most was having a tutor sitting beside us—someone who could watch our work, point out where we went wrong, and give just enough guidance to help us figure it out ourselves. We wanted to recreate that experience with AI. That idea became Socratic Ink: an AI tutor that doesn't just solve problems for students, but actively teaches them by observing their work and adapting to their mistakes.

How we built it

We built Socratic Ink as a closed-loop AI tutoring system that observes each step a student writes on a web-based digital scratchpad instead of only checking the final answer. Students solve physics problems directly in the browser, and when they click Check Step, the backend extracts their latest equation, verifies its correctness, diagnoses the underlying misconception, and chooses the smallest helpful hint rather than revealing the solution. The system then watches the student's next attempt, evaluates whether the intervention worked, updates its misconception model, and adapts its teaching strategy if necessary. We integrated sponsor technologies including Supabase for persistent student state, Nexla for transforming raw interaction events into structured tutoring events, Akash for hosting the tutoring backend, and Zero.xyz for dynamically generating visual explanations whenever verbal hints aren't enough.

Challenges we ran into

One of our biggest challenges was determining the student's actual misconception rather than simply identifying that an answer was incorrect. We also had to design a tutoring loop that knew when to escalate from a small Socratic hint to more direct guidance without immediately giving away the solution. Another challenge was keeping the tutoring experience responsive while coordinating multiple AI components responsible for verification, diagnosis, hint generation, and evaluation. Finally, balancing educational effectiveness with a limited hackathon timeframe required us to narrow our MVP to 1D kinematics while still demonstrating the complete Observe → Verify → Diagnose → Act → Evaluate → Learn loop.

Accomplishments that we're proud of

We're proud that we built more than just another AI homework helper. Instead of generating answers, Socratic Ink actively teaches by observing each step, identifying misconceptions, adapting its hints, and verifying that learning actually occurred through transfer problems. We're also proud of implementing a visible AI reasoning loop that allows users to see how the tutor reaches its decisions, making the system both transparent and educational. Finally, we successfully combined multiple sponsor technologies into a cohesive workflow rather than using them as isolated integrations.

What we learned

Throughout the project, we learned that effective tutoring is far more than providing correct answers—it requires understanding why a student made a mistake and adapting instruction accordingly. We also learned how to build AI systems that continuously evaluate their own interventions instead of stopping after a single response. On the technical side, we gained experience designing event-driven AI workflows, integrating multiple cloud services, and balancing deterministic verification with large language models to create a more reliable educational experience.

What's next for Socratic Ink

Our immediate goal is to expand beyond introductory kinematics into additional STEM subjects such as forces, energy, circuits, calculus, and chemistry. We also plan to improve handwriting recognition so students can write naturally while receiving real-time feedback directly in the browser. Longer term, we'd like to personalize tutoring by learning which teaching strategies work best for each student over time while providing instructors with misconception analytics and progress dashboards. Ultimately, we envision Socratic Ink becoming an AI tutor that provides the personalized guidance of a human tutor while remaining accessible to anyone with an internet connection.

Built With

  • akash
  • nexla
  • python
  • supabase
  • zero.xyz
Share this project:

Updates