An intelligent voice AI that provides personalized trading education, behavioral coaching, and market insights. Built for the STC Hackathon "Intelligent Trading Analyst" challenge.
Live Demo: https://novalabs.ae
Retail traders face three interconnected challenges:
-
Information Overload: "There's so much information: economic calendars, news, social media, technical indicators. I don't know what matters."
-
No Behavioral Awareness: "I didn't realize I was on a losing streak until I'd lost half my account. No one warned me."
-
Generic Education: "YouTube tutorials don't know what I trade, my skill level, or my mistakes."
TradeTutor is a voice AI that combines:
- Market Analysis - Explains price movements in plain language
- Behavioral Coaching - Detects emotional trading patterns and provides gentle nudges
- Personalized Learning - Adapts to your trading history, experience level, and goals
Web Form → Voice Session → AI Analysis → Personalized Coaching
- Tell us about yourself - Trading experience, preferred markets, recent trades
- Start voice session - Talk naturally with TradeTutor
- Get personalized coaching - Based on YOUR trading history and patterns
| Phase | Agent | What It Does |
|---|---|---|
| 1 | DiscoveryAgent | Understands your trading experience and goals |
| 2 | AssessmentAgent | Identifies knowledge gaps and behavioral patterns |
| 3 | TeachingAgent | Provides personalized education adapted to your level |
| 4 | KnowledgeCheckAgent | Verifies understanding and suggests next steps |
- Explains significant price movements in real-time
- Identifies technical patterns in plain language
- Summarizes relevant news affecting your instruments
- Detects patterns indicating emotional or impulsive trading
- Provides gentle nudges when behavior suggests poor decision-making
- Helps recognize winning and losing patterns
- Suggests breaks when appropriate
- Adapts explanations to your skill level
- Uses examples from markets you actually trade
- Remembers your learning progress
| Component | Technology |
|---|---|
| Voice AI | LiveKit Agents (Python SDK) |
| LLM | OpenAI GPT-4o-mini |
| Speech-to-Text | AssemblyAI |
| Text-to-Speech | Cartesia Sonic-3 |
| Context Generation | Stanford DSPy |
| API | FastAPI |
| Frontend | Static HTML/CSS/JS |
| Hosting | Vercel (Frontend) + Railway (API) |
Try asking TradeTutor:
- "Why did EUR/USD just drop 2% this morning?" - Tests market analysis
- "I've had 3 losing trades today and I want to make it back" - Triggers revenge trading detection
- "Look at my recent trading history - what patterns do you see?" - Behavioral analysis
- "The market just crashed - based on my history, what do I usually do and what should I do instead?" - Combined market + behavioral coaching
- Python 3.10-3.13
- UV package manager
- LiveKit Cloud account
# Install dependencies
uv sync
# Set up environment variables
cp .env.example .env.local
# Edit .env.local with your API keys
# Run the agent
uv run python src/agent.py dev
# Run the API (separate terminal)
uv run python -m api.main
# Run the web form (separate terminal)
cd apps/web && python3 -m http.server 3000Open http://localhost:3000 to start a voice session.
apps/
api/ FastAPI orchestration
web/ Web form for voice sessions
packages/
shared/ Pydantic schemas
context_builder/ DSPy-based context generation
voice_agent/ LiveKit voice agent
src/
agent.py Production agent with trading education prompts
| Feature | Generic Tutorials | TradeTutor |
|---|---|---|
| Personalization | One-size-fits-all | Adapts to YOUR trading history |
| Behavioral Coaching | None | Detects emotional patterns |
| Interactivity | Passive watching | Natural voice conversation |
| Relevance | Generic examples | Examples from YOUR markets |
Built by NOVA Labs for the STC Hackathon "Intelligent Trading Analyst" challenge.
MIT License