AI-Powered Concentrated Liquidity Position Manager for Orca Whirlpools on Solana
Maximize your DeFi yields with intelligent range predictions
YieldSense is an advanced liquidity management platform that combines machine learning and real-time analytics to optimize concentrated liquidity positions on Orca Whirlpools. The platform enables users to make data-driven decisions when providing liquidity on the Solana blockchain.
YieldSense is deployed and publicly accessible at:
Live URL: https://yieldsense-app.web.app/
This deployment mirrors the full functionality of the development environment and demonstrates a production-ready architecture for AI-powered DeFi tools.
- Machine learning models analyze historical price data and market volatility
- Dynamic recommendations that adapt to current market sentiment
- Confidence scores to support informed decision-making
- 24-hour yield calculations based on pool volume and fee tier
- Concentration heuristics showing expected returns before deposit
- Accurate fee tier scaling (0.01%, 0.04%, 0.30%)
- Real-time out-of-range notifications when positions require attention
- Firebase-powered monitoring infrastructure
- Customizable alert thresholds per position
- Modern, responsive user interface with glassmorphism design
- Live price charts and liquidity distribution visualization
- Streamlined position creation workflow
+-------------------------------------------------------------------+
| YIELDSENSE STACK |
+-------------------------------------------------------------------+
| |
| +----------------+ +----------------+ +----------------------+ |
| | Frontend | | Backend | | ML API | |
| | (React) | | (Express) | | (FastAPI/Python) | |
| | | | | | | |
| | - Dashboard | | - Position | | - Price Prediction | |
| | - Charts |<>| Manager |<>| - Volatility Model | |
| | - Wallet | | - WebSocket | | - Sentiment Analysis | |
| | Connect | | - Pool Data | | - Staking APY | |
| +----------------+ +----------------+ +----------------------+ |
| | | | |
| +------------------------------------------------------------+ |
| | SOLANA BLOCKCHAIN | |
| | Orca Whirlpools | SPL Tokens | |
| +------------------------------------------------------------+ |
| |
| +----------------+ +---------------------------------------+ |
| | Monitoring | | Firebase | |
| | Service |<>| - Alert Rules - User Preferences | |
| | (Telegram) | | | |
| +----------------+ +---------------------------------------+ |
| |
+-------------------------------------------------------------------+
| Layer | Technology |
|---|---|
| Frontend | React 19, TypeScript, Vite, TailwindCSS, Radix UI |
| Backend | Node.js, Express, TypeScript, WebSocket |
| ML API | Python, FastAPI, scikit-learn, TensorFlow, Transformers |
| Blockchain | Solana, Orca Whirlpools SDK, Anchor Framework |
| Database | Firebase Firestore |
| Alerts | Telegram Bot API |
| Hosting | Thinkroot, Render.com |
- Node.js 18+
- Python 3.10+
- Solana Wallet (Phantom, Solflare, or compatible)
# Clone the repository
git clone https://github.com/Manimaran-tech/Yeildsense.git
cd Yeildsense
# Install dependencies
npm install
cd whirlpool-dashboard && npm install
cd server && npm install
cd ../ml-api && pip install -r requirements.txt# Windows - Launch all services in separate windows
powershell -ExecutionPolicy Bypass -File start_services.ps1Services:
| Service | Port | Description |
|---|---|---|
| Frontend | 3005 | React Dashboard (Vite) |
| Backend | 3001 | Position Manager API |
| ML API | 8000 | AI Prediction Service |
| Trading API | 3002 | Swap Aggregation Service |
| Pool | Fee Tier | Status |
|---|---|---|
| SOL/USDC | 0.01% | Active |
| SOL/USDC | 0.04% | Active |
| JupSOL/SOL | 0.01% | Active |
| SOL/PENGU | 0.30% | Active |
| JUP/SOL | 0.30% | Active |
- Algorithm: Gradient Boosting with LSTM hybrid approach
- Features: OHLCV data, volatility metrics, volume trends
- Performance: Approximately 78% directional accuracy (24-hour horizon)
- Model: GARCH(1,1) for short-term volatility estimation
- Output: Expected price range with confidence intervals
- Sources: Real-time RPC inflation rate and MEV rewards
- Supported Tokens: JupSOL
- Client-Side Signing: No private keys are stored on servers
- Environment Variables: All secrets are managed through environment configuration
- Rate Limiting: Configured RPC endpoints with appropriate limits
MIT License - see LICENSE for details.
YieldSense - Smarter Liquidity, Better Yields