Back to missions
Mobile

Abhyas

A mobile study-generation app that creates chapters, questions, Q&A, and exercises around the topic and depth a learner chooses.

Jun 2024mobileGemini API · Appwrite · Firebase · Android
A

Start with a question

Abhyas began with a simple learning problem: when I want to understand a topic quickly, I do not always need a full course. Sometimes I need a focused explanation, a few chapters, and questions that help me check whether I actually understood it.

The idea was to make that first study session easier to start. A learner could enter any topic, add a little context, and choose how deep the session should go.

Shape the session

The setup keeps the request small but meaningful:

  • A topic name and optional subtopics
  • Basic, intermediate, or expert complexity
  • A chosen number of chapters
  • No exercise, Q&A, or quiz mode
  • A question count when exercises are enabled

Those choices give the generated material a useful direction instead of treating every learner and every subject the same way.

One topic, several ways to learn

Once the study is created, Abhyas turns the response into a small learning space. The Study area presents the generated chapters, Exercises provides a way to test the material, and Notes gathers the longer explanations in a readable format.

The same topic can therefore move from introduction to detail to recall without making the learner assemble those pieces themselves. A session about computer science, for example, can become chapters on networking, a set of questions, and notes explaining how networks work.

The first AI product I explored

I built Abhyas for the Google Gemini API Developer Competition. It was my first serious experience working with AI inside a product, and I was also using Appwrite while learning how authentication, data, and generated content could fit together.

At the time, I was curious whether a focused learning tool could become something larger. I was still learning the foundations of product development, so the project became both an experiment in that idea and a practical lesson in turning an AI response into a usable experience.

What stayed with me

The most important part was not generating more text. It was giving that text a structure that supports a real study loop: choose a direction, read something clear, and then check what stayed with you.