Welcome to Pattern-Signal-Automator – not merely a chart scanner, but an evolving symbiosis between machine vision and human financial intuition. In 2026, the markets have become faster, deeper, and more complex. This tool is your personal cartographer of chaos, translating raw price action into structured, actionable signals.
Inspired by the need for automated pattern recognition that goes beyond simple indicators, this repository delivers a full-stack, cloud-ready system designed to detect, classify, and alert on chart formations (e.g., head and shoulders, triangles, flags, wedges) in real-time or historical data.
Think of it as a neurological bridge: your trading brain meets algorithmic clarity. No more squinting at screens for hours. Let the machine handle the geometry; you handle the strategy.
Traditional chart pattern scanners are often:
- Siloed to one asset class
- Dependent on manual input
- Lacking real-time adaptive learning
Pattern-Signal-Atomator was conceived to solve these exact constraints. It is:
- Multi-asset by design – equities, crypto, forex, futures
- Self-improving – integrates feedback loops for pattern accuracy
- API-native – connect to any data source or broker
“The best patterns are the ones you don’t have to hunt for. Let the algorithm find the needle, while you decide if the haystack is worth burning.”
graph TD
A[Market Data Feed] --> B[Pattern Recognition Engine]
B --> C[Geometric Classifier]
C --> D[Signal Confidence Scoring]
D --> E[Notification Layer]
E --> F[Telegram / Discord / Email]
B --> G[Backtesting Module]
G --> H[Historical Performance DB]
H --> I[Portfolio Simulator]
I --> J[Risk Assessment API]
B --> K[User Feedback Loop]
K --> B
This architecture ensures that every pattern found is not just identified, but contextualized within your broader trading environment.
- Head & Shoulders (Inverse too)
- Ascending / Descending / Symmetrical Triangles
- Bull & Bear Flags / Pennants
- Double / Triple Tops & Bottoms
- Wedges (Rising & Falling)
- Cup & Handle
- Diamond Reversals
Built with React 2026 and Tailwind CSS v4 – works beautifully on desktop, tablet, and mobile landscapes. No more squinting on your phone during a commute.
Supports 14 languages out-of-the-box:
- English, Spanish, French, German, Japanese, Chinese (Simplified), Korean, Portuguese, Arabic, Russian, Hindi, Dutch, Italian, Turkish.
- Websocket-based monitoring across multiple timeframes (1m, 5m, 15m, 1h, 4h, 1D, 1W)
- Configurable alert thresholds
- Cooldown logic to prevent alert fatigue
- OpenAI GPT-5o for natural language summaries of detected patterns
- Claude Opus 4 for alternative interpretation and risk scoring
- Run both models in parallel for a dual-perspective signal validation
- Embedded chatbot (powered by Claude) for technical questions
- Community forum with upvote-based troubleshooting
- Real-time email and ticket system
One-line launch with Docker Compose. No environment conflict, no dependency hell.
Pattern-Signal-Automator/
├── core/ # Pattern detection algorithms
├── api/ # REST & WebSocket endpoints
├── ui/ # Frontend (React + Tailwind)
├── integrations/ # OpenAI, Claude, broker connectors
├── backtesting/ # Historical simulation engines
├── notifications/ # Telegram, Discord, Email, Web push
├── config/ # YAML/JSON config templates
├── tests/ # Unit, integration, and stress tests
├── docs/ # Extended documentation
├── docker-compose.yml # Orchestration
├── Makefile # Build & dev shortcuts
└── README.md # You are here
- Node.js v22+
- Python 3.12+
- Docker & Docker Compose v2.24+
- A free API key from OpenAI and Claude (or run local models)
git clone https://github.com/DevDevahmed/Pattern-Signal-Automator.git
cd Pattern-Signal-Automator
cp .env.example .env # Edit with your API keys
make install # Installs all dependencies
make dev # Launches local development server# config/profiles/trading-energy.yaml
profile_name: "Swing-Low-Hunter"
timeframes: ["1h", "4h", "1D"]
patterns:
- "double_bottom"
- "inverse_head_shoulders"
- "bull_flag"
alert_channels:
- type: "telegram"
token: "your_telegram_bot_token"
chat_id: "-100123456789"
- type: "discord"
webhook_url: "https://discord.com/api/webhooks/..."
parallel_ai: true
openai_model: "gpt-5o-mini"
claude_model: "claude-3.5-opus"# Scan all crypto pairs on Binance for patterns every 5 minutes
python3 automaton.py --exchange binance --symbols ALL_USDT \
--timeframe 5m --patterns triangle,flag,head_shoulders \
--alert telegram --cooldown 1800
# Output example:
[2026-03-15 14:32:01] 🔍 Scanning BTC/USDT (5m)
[2026-03-15 14:32:03] ✅ PATTERN DETECTED: Bull Flag @ $72,450
[2026-03-15 14:32:03] 📊 Confidence: 87.3%
[2026-03-15 14:32:03] 🤖 GPT-5o summary: "Potential continuation pattern with volume confirmation. Target at $74,200 if breakout above flag resistance."
[2026-03-15 14:32:03] 🤖 Claude Opus 4: "Pattern valid but watch for RSI divergence. Slight downside risk if volume drops below average."
[2026-03-15 14:32:04] 📱 Alert sent to Telegram group 'Momentum Traders'| Operating System | Supported | Notes |
|---|---|---|
| 🐧 Ubuntu 24.04+ | ✅ | Primary development target |
| 🪟 Windows 11 | ✅ | WSL2 recommended for Docker |
| 🍎 macOS Sonoma+ | ✅ | M1/M3 optimizations included |
| 🐚 FreeBSD 14 | Limited testing; community support | |
| 🏴☠️ Raspberry Pi OS | ✅ | ARM64 builds available |
| 📦 Docker (any host) | ✅ | Fully containerized mode |
This project naturally integrates critical trading and AI vocabulary without forced repetition. You will find terms such as:
- automated pattern recognition
- real-time market geometry scanning
- AI-assisted technical analysis
- multi-timeframe chart pattern detection
- custom trading signal builder
- open source trading automation
These keywords are embedded in the documentation, configuration files, and code comments so that your search engine visibility grows organically.
When a pattern is detected, the system sends the raw geometric data (coordinates, volume profile, slope, duration) to both LLMs simultaneously. Each returns:
- A natural language summary
- A risk rating (0–10)
- Suggested confirmation triggers (e.g., “Wait for a 2% above breakout”)
These two perspectives are then combined into a Unified Signal Score that gets sent to your alert channels.
{
"parallel_ai": true,
"providers": {
"openai": {
"model": "gpt-5o-mini",
"temperature": 0.3,
"max_tokens": 512
},
"claude": {
"model": "claude-3.5-opus",
"temperature": 0.2,
"max_tokens": 512
}
},
"signal_merger": "weighted_average",
"openai_weight": 0.6,
"claude_weight": 0.4
}This gives you the breadth of GPT and the rigor of Claude in one unified workflow.
This project is released under the MIT License. You are free to use, modify, and distribute this software for commercial or personal projects.
Risk Warning: Automated trading tools, including pattern detection algorithms, are not guaranteed to produce profitable outcomes. Past performance does not indicate future results. Financial markets carry inherent risk of loss.
This software is provided for educational and research purposes only. It is not financial advice, and you should consult a licensed professional before making any investment decisions.
The developers assume no liability for any financial losses incurred through the use of this tool.
“Even the sharpest scalpel cannot perform the surgery on its own. You are the surgeon – don’t rely solely on the blade.”
In a world where 90% of traders fail, the difference between success and burnout often comes down to mental bandwidth. Pattern-Signal-Automator exists to reclaim those hours of staring at indeterminate shapes on a screen. Let the machine do the geometry. You focus on the strategy, the psychology, and the edge that only a human can bring.
Welcome to smarter trading in 2026.