Inspiration
In every classroom, there’s always that one moment — when a professor explains a complex idea so clearly that everything finally clicks. But just as quickly, that moment can fade. Students struggle to take perfect notes while listening, and teachers juggle between engaging their class and managing slides, polls, and materials. We wanted to fix that. Lectra-AI was born from the dream of making lectures smarter, simpler, and more human. What if AI could capture every word, every insight, and every discussion — turning them into beautifully structured notes in real time? What if teachers could teach freely, without worrying about recording, typing, or organizing? That’s the world Lectra-AI is creating.
What it does
Lectra-AI transforms ordinary lectures into intelligent, interactive experiences. It listens to the teacher’s voice through Deepgram’s Speech-to-Text, converts it into text, and uses Gemini AI to create clean, readable bullet-point notes — updated live on a digital “blackboard” for every student to see. Teachers can even control Lectra hands-free through voice commands like “AI, open a new slide” or “AI, start a quiz.” The system instantly responds — generating polls or quiz questions that students can answer in real time. At the end of each session, Lectra compiles everything — from notes to quizzes — into a beautifully formatted PDF summary, making learning more accessible and review-friendly than ever.
How we built it
We built Lectra-AI with a modern full-stack architecture designed for real-time collaboration: Frontend: React.js + Tailwind CSS for a sleek, responsive “live blackboard” experience. Backend: Node.js + Express + Socket.IO to enable real-time data flow between teachers and students. AI Stack: Deepgram handles the live transcription, while Gemini AI generates clean, structured notes. Database: MongoDB stores lectures, users, and generated summaries. PDF Engine: Puppeteer converts AI summaries into downloadable lecture PDFs. Authentication: JWT ensures secure, role-based access for teachers and students. Together, these technologies form a seamless, low-latency pipeline that feels effortless to use — like magic in motion.
Challenges we ran into
Building real-time voice-driven interaction was both thrilling and tricky. Handling audio streaming latency while ensuring smooth transcription updates required careful optimization. Designing a Socket.IO system that could support multiple students viewing the same live notes was a challenge in synchronization. Integrating Gemini’s generative capabilities into real-time flows demanded precise prompt design to avoid lag or over-generation. Balancing performance, scalability, and accuracy was a continuous process — but it made the system stronger every step of the way.
Accomplishments that we're proud of
Achieved real-time AI note generation from live speech with minimal delay. Built a fully functional voice command system that feels natural and responsive. Designed an intuitive dual-interface system — a live lecture page for teachers and a dynamic blackboard for students. Created automatic summary and quiz PDF generation, turning every class into reusable learning material. Most importantly, we made something that genuinely helps both teachers and students — a tool that reimagines the modern classroom.
What we learned
Throughout this journey, we learned that building with AI isn’t just about automation — it’s about augmentation. When technology amplifies human communication instead of replacing it, the results are powerful. We also deepened our understanding of: Streaming architectures and real-time WebSocket communication. Prompt tuning for consistent AI output in live settings. Designing interfaces that are both educational and emotionally engaging. And above all, we learned that the best innovations are born from empathy — understanding the challenges teachers and students face every day.
What's next for Lectra-AI
Lectra-AI is just getting started. Next, we plan to: Add multi-language support for international classrooms. Integrate video recording and playback alongside transcripts. Build a mobile companion app for students to review notes on the go. Introduce advanced analytics dashboards for teachers to track engagement and quiz performance. Explore LMS integrations with platforms like Canvas and Google Classroom. Our vision is clear — to make Lectra-AI the standard assistant in every classroom, where learning is not limited by typing speed, attention span, or missed moments — but powered by understanding.
Built With
- axios
- bcrypt.js
- cors
- css
- deepgram
- express.js
- geminiapi
- javascript
- jwt
- mongodb
- multer
- node.js
- puppeteer
- react.js
- socket.io
- tailwind

Log in or sign up for Devpost to join the conversation.