IMPORTANT DISCLAIMER: This software is for educational and research purposes only. It is NOT financial advice and should NOT be used for actual trading or investment decisions. You are solely responsible for your investment decisions and any financial losses.
CryptVault is an advanced cryptocurrency and stock analysis platform featuring a native desktop GUI, production-grade machine learning predictions, comprehensive technical analysis, and sophisticated pattern detection.
- Native Desktop App: Dark-themed trading terminal with candlestick charts, pattern overlays, and ML predictions
- Pattern Detection: 50+ patterns across 7 categories — candlestick, reversal, continuation, harmonic, and divergence
- Actual Pattern Drawing: Reversal and continuation patterns are drawn as geometric shapes on the chart (necklines, trendlines, pivot connections)
- Production ML System: Ensemble predictor with stacking, 67+ engineered features, and validation-based weighting
- Reinforcement Learning: RL agents (DQN, PPO, Transformer) for trading research
- Technical Indicators: Bollinger Bands, RSI, Volume — cleanly rendered without clutter
- Multi-Asset Support: Cryptocurrencies and stocks via Yahoo Finance
- Average MAPE: 1.6-2.4% across major cryptocurrencies
- Direction Accuracy: 100% on tested symbols
- Within 2% Accuracy: 80-100% of predictions
- R2 Score: 0.50-0.81
Tested on BTC, ETH, SOL, BNB with 120 days of historical data.
- Desktop App Overhaul: Removed cluttered moving averages, removed unnecessary real-time streaming tab
- Actual Pattern Drawing: Double Tops/Bottoms, Head & Shoulders, Triangles, and Wedges now draw actual geometric shapes (trendlines, necklines, pivot connections) on the chart instead of just markers
- Fixed Pattern Detection: Double Bottom strength calculation bug fixed; all chart patterns now store key point coordinates in
extrafield - Cleaner Analysis: Analysis panel always visible alongside chart in a split pane
- Simplified Navigation: Chart and Analysis only — no unused real-time tab
- Docs: Added DESKTOP_APP.md and PATTERNS.md
Developed by MeridianAlgo - Algorithmic trading research and development.
git clone https://github.com/MeridianAlgo/Cryptvault.git
cd Cryptvault
pip install -r requirements.txtpython launch_desktop.py# Analyze Bitcoin with chart
python cryptvault_cli.py BTC 60 1d
# Save chart to file
python cryptvault_cli.py ETH 120 1d --save-chart eth_chart.png
# Text-only analysis
python cryptvault_cli.py SOL 90 1d --no-chartThe desktop app provides a full trading terminal experience.
+------------------------------------------------------------------+
| CryptVault v6.0 [Symbol: BTC-USD] [Analyze] 1D 5D 1M 3M ... |
+----------+--------------------------------------+------------------+
| Chart | | ML Prediction |
| Analysis | Candlestick Chart | ──────────── |
| | Bollinger Bands | Detected |
| ─────── | Pattern shapes (drawn) | Patterns |
| LAST | Volume | (scrollable) |
| ANALYSIS | RSI | |
+----------+--------------------------------------+------------------+
| Status bar |
+------------------------------------------------------------------+
Patterns are rendered as actual geometric shapes:
| Pattern | Drawn As |
|---|---|
| Double Top / Bottom | Lines connecting peaks/troughs + dashed neckline |
| Head & Shoulders | LS-Head-RS connection + dashed neckline |
| Triple Top / Bottom | Lines connecting all three pivot points |
| Triangles (Sym/Asc/Desc) | Upper and lower fitted trendlines |
| Rising / Falling Wedge | Upper and lower fitted trendlines |
| All patterns with target | Dotted horizontal target price line |
| Candlestick patterns | Triangle marker (^ bullish, v bearish) |
See Desktop App Documentation for full details.
50+ patterns across 7 categories. See Pattern Reference.
Reversal Patterns
- Head and Shoulders / Inverse Head and Shoulders
- Double Top / Double Bottom
- Triple Top / Triple Bottom
- Rising Wedge / Falling Wedge
Continuation Patterns
- Ascending Triangle / Descending Triangle / Symmetrical Triangle
- Bull Flag / Bear Flag / Pennants
- Cup and Handle
Candlestick Patterns
- Doji (Standard, Gravestone, Dragonfly)
- Hammer, Hanging Man, Inverted Hammer, Shooting Star
- Engulfing, Harami, Piercing Line, Dark Cloud Cover
- Morning Star, Evening Star, Three White Soldiers, Three Black Crows
Harmonic Patterns
- Gartley, Butterfly, Bat, Crab, Shark, Cypher
Divergence
- RSI Bullish/Bearish Divergence
- MACD Bullish/Bearish Divergence
- Random Forest: Tree-based ensemble learning
- Gradient Boosting: Sequential model optimization
- Support Vector Machines: Non-linear regression
- Linear Models: Ridge, Lasso, ElasticNet
- ARIMA: Time series forecasting
- XGBoost/LightGBM: Advanced gradient boosting (optional)
Each model is weighted by historical validation accuracy.
State-of-the-art RL agents for trading research:
- DQN: Deep Q-Network with Rainbow improvements (dueling, noisy nets, prioritized replay)
- PPO: Proximal Policy Optimization with GAE
- Transformer: Multi-head attention for sequential decisions
See RL Documentation for usage.
python cryptvault_cli.py BTC 60 1d
python cryptvault_cli.py ETH 90 1d --save-chart ethereum_analysis.png
python cryptvault_cli.py SOL 30 4hpython cryptvault_cli.py AAPL 60 1d
python cryptvault_cli.py TSLA 90 1d --save-chart tesla.png
python cryptvault_cli.py GOOGL 60 1d --verbosepython cryptvault_cli.py --portfolio BTC:0.5 ETH:10
python cryptvault_cli.py --compare BTC ETH SOL
python cryptvault_cli.py --interactive
python cryptvault_cli.py --statuspython cryptvault_cli.py SYMBOL [DAYS] [INTERVAL] [OPTIONS]| Option | Description |
|---|---|
--no-chart |
Text-only output |
--save-chart FILE |
Save chart image |
--verbose |
Detailed output |
--demo |
Run demonstration |
--desktop |
Launch desktop app |
--version |
Show version |
--help |
Show help |
--portfolio ASSET:AMT ... |
Portfolio analysis |
--compare SYM1 SYM2 ... |
Compare assets |
--interactive |
Interactive REPL |
--status |
Check API status |
CryptVault/
cryptvault/
desktop/ Native desktop GUI
app.py
main_window.py
theme.py
panels/
chart_panel.py
analysis_panel.py
patterns/ Pattern detection
comprehensive.py (50+ patterns, 7 categories)
ml/ Machine learning models
rl/ Reinforcement learning agents
data/ Data fetching and management
visualization/ Chart generation
portfolio/ Portfolio analysis
docs/
DESKTOP_APP.md Desktop app guide
PATTERNS.md Pattern detection reference
ARCHITECTURE.md
API_REFERENCE.md
TROUBLESHOOTING.md
tests/
cryptvault_cli.py CLI entry point
launch_desktop.py Desktop launcher
requirements.txt
- Python 3.9+
- 4 GB RAM
- 2 GB disk space
- Internet connection (data fetching)
- Python 3.11+
- 8 GB RAM
- 5 GB disk space
- Windows 10/11
- Ubuntu 20.04+
- macOS 10.15+ (including Apple Silicon)
pytest tests/ -v
pytest tests/ -v --cov=cryptvault --cov-report=htmlblack cryptvault/ cryptvault_cli.py
isort cryptvault/ cryptvault_cli.py
flake8 cryptvault/
mypy cryptvault/
bandit -r cryptvault/- Desktop App Guide
- Pattern Reference
- Architecture Overview
- API Reference
- Changelog
- Troubleshooting
- Contributing Guide
- Security Policy
Read CONTRIBUTING.md and CODE_OF_CONDUCT.md before submitting pull requests.
- Fork the repository
- Create a feature branch
- Install dev dependencies:
pip install -r requirements.txt - Make your changes with tests
- Run tests and linting
- Submit a pull request
MIT License. See LICENSE.
- MeridianAlgo Team: Core development
- scikit-learn: Machine learning
- yfinance: Market data
- NumPy, pandas, SciPy: Scientific computing
- Matplotlib: Visualization
- XGBoost, LightGBM: Gradient boosting
This software is strictly for educational and research purposes.
- NOT FINANCIAL ADVICE: Does not provide investment or trading advice
- NOT FOR TRADING: Do not use for actual investment decisions
- NO GUARANTEES: Past performance does not guarantee future results
MeridianAlgo is a nonprofit research organization focused on ML research and open-source financial technology. We are not a licensed financial advisor, broker, or investment firm.
Version: 6.0.0 Last Updated: March 2026 Maintained by: MeridianAlgo