AhaLoop

AhaLoop turns student explanations into thought maps, Socratic questions, and measurable learning progress—without simply giving away the answer.

Comment

Inspiration

Most AI learning tools give students answers. But receiving an answer does not prove that a student truly understands the subject.

AhaLoop was created around one question: Did I really understand this—or does it only feel familiar?

What it does

AhaLoop is an AI-powered learning coach. The student chooses an educational level and a topic, then explains what they currently understand in their own words.

GPT-5.6 analyzes the explanation and creates a thought map showing strong connections, missing links, possible misconceptions, and one focused Socratic question.

Instead of giving away the answer, AhaLoop asks the student to reflect and explain again. After three focused rounds, it generates a Mastery Card showing the student's progress, concepts mastered, changes in thinking, topics to review, a clear mental model, and a final transfer challenge.

The Mastery Card can also be printed or saved as a one-page PDF. AhaLoop works in English and Spanish and does not require an account.

How we built it

AhaLoop was built from scratch during OpenAI Build Week with Codex and GPT-5.6.

The application uses Next.js, React, TypeScript, Tailwind CSS, Framer Motion, the OpenAI Responses API, Structured Outputs with Zod, local browser storage, and Vercel.

GPT-5.6 powers the structured educational analysis and generates the Socratic question for each learning round. The OpenAI API key and internal instructions remain securely on the server.

Codex collaborated throughout product planning, architecture, implementation, responsive design, API integration, testing, debugging, documentation, and deployment.

Challenges we ran into

Our main challenge was creating an AI learning experience that guides students without becoming an answer machine. We also needed consistent structured responses, a clear three-round experience, bilingual support, secure error handling, responsive mobile layouts, and a printable one-page Mastery Card.

Accomplishments that we're proud of

AhaLoop transforms an imperfect explanation into a visible learning journey.

In our demonstration, a student begins with the misconception that seasons are caused by Earth's changing distance from the Sun. Through Socratic guidance, the student connects Earth's axial tilt, sunlight angle, daylight length, and temperature.

The demonstrated progression is 25% → 78% → 96%.

What we learned

We learned that educational AI becomes more valuable when it analyzes reasoning instead of simply evaluating a final answer. One carefully selected question can help a student build a stronger mental model while keeping the intellectual work with the learner.

What's next for AhaLoop

Future versions could include optional student and teacher accounts, classroom dashboards, curriculum-aligned activities, progress across multiple topics, voice explanations, accessibility improvements, and additional languages.

Built With

Share this project:

Updates