What is Medhavy?

Most students already have access to a general-purpose AI chatbot that can answer almost anything. That's actually the problem it creates in a learning context: an answer that isn't tied to what you're supposed to be studying doesn't necessarily help you understand your actual coursework. Medhavy starts from that specific gap and asks a different question: what if your textbook could teach you back?

An AI tutor grounded in the book you're reading

Medhavy is an AI-powered learning platform built around interactive textbooks with an AI tutor embedded directly into the reading experience. The core design decision is that the tutor pulls its answers from the content of the specific textbook a student is using, rather than from a general pool of information. If you're working through a chapter and something doesn't make sense, the tutor stays connected to that material instead of wandering off into a broader, less relevant explanation. The goal isn't just to hand over an answer; it's to help you understand the material you're actually studying.

One login gives students access to whichever books they're authorized to use. Right now those include physics, cancer biology, quantum mechanics, and electron microscopy. Instead of treating the textbook, the tutor, and the learning tools as separate pieces a student has to juggle, Medhavy brings all of it into one environment.

Adapting to how each student actually learns

Not every student learns the same way, and Medhavy is designed around that fact rather than around a single fixed tutoring script. One student might understand something immediately after seeing a visual. Another might need to work through an idea step by step. Someone else might learn best when they're asked the right question and guided toward the answer rather than told it outright. Medhavy is built to remember these interactions and build a learner profile over time, so as a student reads, asks questions, and works through material, the tutor develops a clearer picture of how that particular student learns and where they need more support. Instead of one tutoring experience for everyone, the experience adapts to the learner.

An AI co-instructor, not just a student tool

Medhavy isn't designed only for students. The same learning environment is built to support instructors as well. Instructors already carry a large amount of course knowledge in their textbooks, lecture notes, and other materials, and Medhavy is designed to help them make more use of that content. Course material can be used to generate summaries, quizzes, exams, and lecture notes. The idea is an AI co-instructor that reduces repetitive preparation work while keeping the instructor at the center of the course, rather than replacing their judgment.

Building a course from scratch takes real time: organizing material, creating modules, writing assessments, and preparing supporting resources can add hours of work before a single student even enters the classroom. Medhavy is designed to shorten that process too. Material can come directly from a textbook or even a GitHub repository and then be organized into course modules with generated assessments, so instructors start from what they already have rather than rebuilding everything by hand.

What it looks like for a student

In practice, a student logs in once and sees the textbooks available to them. They open a book, choose what to study, and start reading much like they would in any digital textbook. The difference is that the learning tools sit right there with the content. If something doesn't make sense, the student doesn't have to leave the book, open a separate tool, and re-explain all the context. They can ask the tutor while they're still reading. And instead of jumping straight to a final answer, the tutor can nudge with hints, explanations, and questions, closer to how a human tutor guides someone through a difficult idea.

The textbook itself is interactive as well. In the quantum mechanics book, for example, students can interact with simulations directly on the page, exploring how changing a variable affects the outcome rather than only reading a static description of it.

Key takeaways

• Medhavy embeds an AI tutor directly inside interactive textbooks, currently covering physics, cancer biology, quantum mechanics, and electron microscopy.
• The tutor answers from the specific textbook a student is using, not from general knowledge, keeping it connected to what's actually being studied.
• It builds a learner profile over time, adapting its explanations, visual, step-by-step, or question-guided, to how each student learns.
• The tutor favors hints and guided questions over immediately handing over the final answer.
• Instructors can use their own textbooks, lecture notes, and materials, or even a GitHub repository, to generate summaries, quizzes, exams, and course modules.
• Interactive elements, like simulations in the quantum mechanics book, let students explore concepts directly on the page.

Who this is for

Medhavy is built for students who want tutoring that stays tied to their actual coursework instead of a generic chatbot answer, and for instructors who want to turn existing course material into a structured learning experience without rebuilding it from scratch. It's a project from the Humanitarians AI ecosystem aimed at making both learning and teaching more efficient and more personalized.