StudyMate

• 

  StudyMate interface

💡 Inspiration

University students are overwhelmed with long lectures and dense slides. We built StudyMate to meet them where they already are — on Telegram — with zero setup. A study assistant that can listen to lectures and read slides for you.

🤖 What It Does

StudyMate is a Telegram AI study assistant with four modes:

• 💬 Conversational — Ask questions or upload course PDFs for context-aware answers (RAG).
• 📖 Read Slide — Upload slides and get them as voice messages.
• 🎙️ Audio to Notes — Convert lectures/voice notes into structured notes.
• 📚 Course Advising — Get personalised academic guidance.

No downloads. No sign-up. Just Telegram.

🛠️ How We Built It

• Backend: FastAPI (Telegram webhooks)
• Bot: Telegram Bot API
• AI: Groq (LLM inference)
• Speech: Azure Speech (STT + TTS)
• RAG: Vector DB for course-specific Q&A

🚧 Challenges

• Handling PDFs, audio, and voice seamlessly
  
• Meeting Telegram’s 5s webhook limit with heavy tasks
  
• Cleaning slide text for natural speech output
  
• Improving RAG relevance
  

🏆 Accomplishments

• Fully functional AI assistant inside Telegram
  
• Real-time voice → text + voice responses
  
• Personalised answers via RAG
  
• Modular, scalable architecture
  

📚 What We Learned

• Building multi-modal Telegram bots
  
• Integrating LLMs, speech AI, and RAG
  
• Importance of background processing for performance
  

Built With

• azure-speech-services
• fastapi
• groq
• llm
• python
• rag
• telegram-bot-api
• vector-database