Inspiration

Every learner deserves a patient tutor who can explain hard ideas out loud, adapt to their pace, and keep them moving even when a teacher or parent is not available. We built Teachnow AI around Lumi, a voice-based AI tutor for any subject, because learning often works best as a conversation: asking questions, trying examples, getting unstuck, and building confidence step by step.

What it does

Teachnow AI helps learners take real courses with Lumi as their personal tutor. Learners can talk through concepts, practice problems, review notes, follow structured curricula, and resume progress from their dashboard. The platform also supports community courses, including richer K-12 curricula such as CBSE Class 11 Mathematics for India learners, plus study circles where instructors can run cohort-based learning with sessions, materials, homework, announcements, discussions, and recordings.

How we built it

We built Teachnow AI as a full learning platform with a polished learner dashboard, structured course runtime, voice tutoring, curriculum navigation, community catalog, and study circle workflows. A major hackathon milestone was PR 570, where we overhauled the buggy realtime tutor engine. Instead of relying on a fragile realtime flow, the app now commits the next learning card first and lets Lumi speak after the learner-facing state is ready. This made the tutoring experience much more reliable and predictable.

This week, we also expanded study circles and generated richer India-focused K-12 curriculum content using GPT 5.6 Sol.

Challenges we ran into

The hardest part was making voice tutoring feel reliable, not just impressive in a demo. Realtime AI can easily become inconsistent if the tutor speaks before the app has committed the next card, exercise, or curriculum state. We had to redesign the engine around deterministic learning state so the UI, curriculum, and voice experience stay in sync.

Another challenge was making Teachnow AI work for both self-paced learners and cohort-based classes without turning the product into two separate systems.

Accomplishments that we're proud of

We are proud that Teachnow AI now feels like a real learning product, not just a chatbot. Learners can see their progress, resume real courses, move through structured curricula, and talk with Lumi naturally.

We are also proud of the study circle work. Instructors now have real course hubs with enrollment controls, capacity editing, live session prep, saved materials, homework, announcements, discussion, progress, calendar export, and review workflows. Students get a cohort experience around the same learning content.

What we learned

We learned that the quality of an AI learning experience depends as much on product state and curriculum structure as on the model itself. A great tutor needs to know what the learner is seeing, what came before, what should come next, and how to recover cleanly when something changes.

We also learned that voice-first learning needs careful pacing. The AI should not just answer; it should guide, pause, check understanding, and keep the learner oriented.

What's next for Teachnow AI

Lumi, our voice-based AI tutor for any subject, is personalized to each learner. Teachnow AI helps anyone talk through hard ideas, move at the right pace, and keep learning beyond what fits inside a single classroom period.

Next, we want to make Lumi even more adaptive across subjects, deepen curriculum quality for more countries and school systems, improve study circle tools for instructors, and make Teachnow AI easier for schools, parents, and independent learners to use every day.

Built With

Share this project:

Updates