Your voice-powered financial guide — simplifying mutual fund investments through natural conversation.
Millions of beginners find mutual funds intimidating — jargon, endless choices, and complex data create a barrier to informed investing.
Advisor breaks that barrier. It's a voice-first AI assistant that understands natural language, adaptively profiles your risk appetite, analyzes real market data, and generates a personalized investment report — all through a conversation that feels human.
Who it helps: Young professionals exploring their first investments, families seeking simplified guidance, beginners wanting to learn without the jargon, and anyone who prefers speaking over filling out forms.
| Category | Features |
|---|---|
| 🎙️ Voice | Speech-to-Text, Text-to-Speech, real-time transcription, noise handling |
| 🧠 AI | Natural language understanding via Ollama + Gemma 3, conversation memory, adaptive questioning, risk classification |
| 📊 Analytics | CAGR, volatility, Sharpe ratio, maximum drawdown, historical NAV (10+ yrs) |
| 🔒 Security | Input validation, rate limiting, centralized error handling, env isolation |
| 🚧 Future | Multi-language, portfolio simulation, fund comparison, SIP calculator |
| 🔗 Live Demo | 🎥 Demo Video | 📊 Presentation |
| Coming Soon | Coming Soon | Coming Soon |
| Category | Technology |
|---|---|
| Frontend | React 18 + Vite · Web Speech API |
| Backend | Node.js + Express · MongoDB + Mongoose |
| AI / LLM | Ollama + Gemma 3 (local) |
| Analytics | Python + FastAPI · pandas + NumPy |
| Data Source | mfapi.in (Mutual Fund Data) |
graph TB
User -->|Voice| WS[Web Speech API<br/>STT / TTS]
WS -->|Text| FE[Frontend<br/>React + Vite]
FE -->|API| BE[Backend<br/>Node.js + Express]
BE -->|LLM Prompt| OLLAMA[Ollama · Gemma 3]
BE -->|Fetch & Analyze| ANALYTICS[Analytics<br/>Python + FastAPI]
ANALYTICS -->|NAV Data| MFAPI[mfapi.in]
BE -->|Store| DB[(MongoDB)]
BE -->|Report| FE
Flow: User speaks → Speech-to-Text → LLM risk analysis → Fetch real fund data → Compute metrics → Generate personalized report → Display in frontend.
Make sure the following are installed on your machine:
- Node.js 18+ and npm
- Python 3.10+
- Ollama installed and available in your PATH
- MongoDB running locally or a MongoDB Atlas connection string
git clone https://github.com/HARSHILL2023/codeavengers.git
cd advisor-project# Backend
cd advisor-backend
npm install
Copy-Item .env.example .env
# Frontend
cd ../advisor-frontend
npm install
Copy-Item .env.example .env
# Analytics
cd ../advisor-analytics
python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txtIf PowerShell blocks script activation, run:
Set-ExecutionPolicy -Scope Process Bypassollama pull llama3.2:3bOpen separate terminals and run the following commands.
cd advisor-project\advisor-analytics
.\.venv\Scripts\Activate.ps1
uvicorn main:app --reload --port 8000cd advisor-project\advisor-backend
ollama serve
node server.jscd advisor-project\advisor-frontend
npm run devYou can also start everything from the repo root with:
.\start-all.ps1Backend environment variables in advisor-backend/.env.example:
PORT=5000
MONGODB_URI=mongodb://127.0.0.1:27017/advisorDB
ANALYTICS_SERVICE_URL=http://localhost:8000
OLLAMA_URL=http://localhost:11434
OLLAMA_MODEL=llama3.2:3bFrontend environment variables in advisor-frontend/.env.example:
VITE_API_URL=http://localhost:5000/api
VITE_SUPABASE_URL=https://your-project.supabase.co
VITE_SUPABASE_ANON_KEY=your-anon-key| Service | Variable | Description | Required |
|---|---|---|---|
| Backend | PORT |
Backend port | ❌ |
| Backend | MONGODB_URI |
MongoDB connection string | ✅ |
| Backend | ANALYTICS_SERVICE_URL |
Analytics service base URL | ❌ |
| Backend | OLLAMA_URL |
Ollama base URL | ❌ |
| Backend | OLLAMA_MODEL |
Ollama model name | ❌ |
| Frontend | VITE_API_URL |
Backend API URL | ❌ |
| Frontend | VITE_SUPABASE_URL |
Supabase project URL | ❌ |
| Frontend | VITE_SUPABASE_ANON_KEY |
Supabase anonymous key | ❌ |
- Open
http://localhost:5173in Chrome/Edge - Click the microphone and start speaking
- The AI asks about your financial goals, income, and timeline
- Your risk profile is classified as Conservative 🟢, Moderate 🟡, or Aggressive 🔴
- Real fund data is fetched matching your profile
- A personalized investment report is generated with CAGR, volatility, Sharpe ratio, and recommendations
Landing |
Voice |
Chat |
Report |
| Step | Description |
|---|---|
| Speech → Text | Web Speech API captures audio and transcribes in-browser — zero external STT services |
| Conversation Memory | Full session history stored in MongoDB, fed to LLM for coherent multi-turn dialogue |
| Adaptive Questioning | LLM assesses information gaps and dynamically generates follow-up questions |
| Risk Classification | Structured prompting classifies users into Conservative / Moderate / Aggressive |
| Analytics Engine | Python service fetches NAV data via mfapi.in, computes CAGR/volatility/Sharpe/drawdown with pandas |
| Report Generation | LLM receives analytics data and produces a human-readable investment report |
| Phase | Features |
|---|---|
| ✅ v1.0 | Voice conversation, adaptive risk profiling, real fund data, personalized reports, financial analytics |
| 🚧 v1.5 | Multi-language support, fund comparison, portfolio simulation, PDF export |
| 🔮 v2.0 | User accounts, email/SMS delivery, SIP calculator, mobile app, real-time market data |
| Name | GitHub |
|---|---|
| Harshil Patel | @HARSHILL20233 |
| Anand Suthar | @anand880441-source |
| Trikum Devasi | @TrikamDevasi |
| Pritesh Bachhav | @BachhavPritesh |
MIT — see LICENSE.
Hackathon organizers · mfapi.in (free mutual fund data) · Ollama (local LLMs) · React + Vite · FastAPI + pandas
If you found this helpful, give it a ⭐ on GitHub!
Built with ❤️ for hackathon