An AI-powered NCERT tutoring chatbot that explains concepts in Hinglish (Hindi + English) and Tanglish (Tamil + English) — the way students actually talk.
Built with Streamlit + Groq AI (LLaMA 3.3 70B).
Live app: studybuddy-ai-ncert.streamlit.app
- AI tutor with a friendly college-senior personality
- Natural Hinglish & Tanglish code-switching (not robotic translation)
- Context-aware responses tied to class, subject, and chapter
- Conversation memory within sessions
- Classes 8, 9, 10: Science, Social Science, Mathematics, English, Hindi, Tamil
- Classes 11, 12: Three streams — Bio-Maths, Computer Science, Commerce
- Bio-Maths: Physics, Chemistry, Biology, Mathematics, English
- Computer Science: Physics, Chemistry, Mathematics, CS, English
- Commerce: Business Studies, Economics, Accountancy, Mathematics, English
- Full NCERT chapter lists for every class/subject combination
- Tamil follows TN Samacheer Kalvi syllabus
- Notes — AI-generated revision cheat sheets per chapter
- Quiz — 5 MCQs with scoring, explanations, and retry
- Flashcards — 8 flip cards with shuffle and navigation
- All tools regenerate when you switch language or chapter
- Create unlimited chat sessions with "New Chat"
- Auto-titled from first message
- Switch between past sessions instantly
- Persistent JSON storage across page reloads
- Clear individual or all history
- Floating mic button using Web Speech API (Chrome)
- Real-time speech-to-text in en-IN locale
- Visual recording indicator
- Images (PNG, JPG) — analyzed by LLaMA 4 Scout vision model
- PDFs — text extracted and used as context for Q&A
- Upload diagrams, textbook pages, or handwritten problems
- Animated mascot in header (CSS keyframe study-bob animation)
- Purple gradient theme with light/dark mode support
- Chapter badge showing Class > Subject > Chapter
- Starter questions per chapter for quick start
- Responsive centered layout
| Component | Technology |
|---|---|
| Frontend | Streamlit 1.40+ |
| AI Model | Groq API — llama-3.3-70b-versatile |
| Vision Model | meta-llama/llama-4-scout-17b-16e-instruct |
| PDF Parsing | pypdf 4.0+ (optional) |
| Voice | Web Speech API (browser-native) |
| Storage | JSON file (chat_sessions.json) |
| Language | Python 3.x |
studybuddy-ai/
app.py Main Streamlit application (~1100 lines)
config.py Classes, subjects, chapters, starter questions
prompts.py Hinglish & Tanglish system prompts
mascot.png Animated header mascot
requirements.txt Python dependencies
.env GROQ_API_KEY (not committed)
chat_sessions.json Persistent chat history (auto-generated)
-
Clone the repo
git clone https://github.com/nareshh3105/srm-hackathon.git cd srm-hackathon -
Install dependencies
pip install -r requirements.txt
-
Add your Groq API key
Create a
.envfile in the project root:GROQ_API_KEY=your_groq_api_key_hereGet a free key at console.groq.com
-
Run the app
streamlit run app.py
-
Open
http://localhost:8501in your browser.
The AI doesn't translate — it thinks natively in the target language:
Hinglish example:
Dekh bhai, photosynthesis basically plants ka apna khana khud banana hai! Sunlight uska gas stove hai, CO2 aur paani raw ingredients hain.
Tanglish example:
Seri paaru, photosynthesis-nu sonna plants thaniyaa saapadu panni'kkum process! Sunlight adoda gas stove, CO2-um water-um raw ingredients.
Technical terms (photosynthesis, voltage, democracy, polynomial) always stay in English — never translated.
Class selector (8-12)
+-- Group selector (only for 11/12: Bio-Maths / CS / Commerce)
+-- Subject selector (dynamic per class/group)
+-- Chapter selector (full NCERT chapter list)
+-- Chat / Notes / Quiz / Flashcards
- Switching class, group, or subject auto-saves the current chat and starts a new session
- Switching language regenerates the last AI response and all tool content
- Chat history persists in
chat_sessions.jsonacross browser refreshes
User Input (text / voice / image / PDF)
|
v
app.py (Streamlit UI + session state)
|
v
prompts.py (builds system prompt with class/subject/chapter/language)
|
v
config.py (resolves chapters, starter questions, subject lists)
|
v
Groq API (LLaMA 3.3 70B / LLaMA 4 Scout for vision)
|
v
Response rendered in chat / notes / quiz / flashcards
| Function | File | Purpose |
|---|---|---|
get_response() |
app.py | Main chat API call with context |
get_vision_response() |
app.py | Image analysis via vision model |
generate_quiz() |
app.py | Create 5 MCQs as JSON |
generate_flashcards() |
app.py | Create 8 flashcards as JSON |
generate_notes() |
app.py | Create chapter revision notes |
get_system_prompt() |
prompts.py | Build language-aware system prompt |
get_subjects() |
config.py | Get subject list for class/group |
get_chapters() |
config.py | Get chapter list for class/subject |
get_starter_questions() |
config.py | Get 3 starter questions per chapter |
sync_current_session() |
app.py | Save current session to disk |
load_sessions_data() |
app.py | Load all sessions with migration |
| Stage | What was done |
|---|---|
| v1 — Base | Streamlit + Groq chatbot with Class 10 Science only |
| v2 — Memory | Persistent chat history using chat_history.json |
| v3 — Subjects | Expanded to 6 subjects: Science, Social Science, Maths, English, Hindi, Tamil |
| v4 — CSS | Light/dark mode with proper contrast, gradient theme |
| v5 — Sessions | Multi-session chat with New Chat button, history panel, clear history |
| v6 — Tools | Added Notes, Quiz, and Flashcards generation tools |
| v7 — Classes | Expanded from Class 10 to Classes 8-12 with stream/group support for 11/12 |
| v8 — Rebrand | Renamed to Studybuddy AI, added animated mascot, removed all emojis |
| v9 — Prompts | Enhanced Hinglish & Tanglish prompts for authentic, natural language |
- Sujitha C B (II, CSE, SSN College Of Engineering)
Built for SRM Hackathon
This project is for educational purposes.