AETHER SCAN

• 

  flow chart of this project

Inspiration The inspiration for AETHER SCAN came from watching professional traders spend hours staring at screens, trying to "feel" the market through complex visual patterns. I realized that humans often make mistakes due to emotional bias and fatigue. Being 14 years old, I wanted to build something futuristic—a system that could "see" the market exactly like a human eye does, but with the cold, calculated precision of a machine. My goal was to create a digital co-pilot that never blinks and never sleeps.

What it does AETHER SCAN is a multimodal AI-driven market analyst that performs live pixel analysis. Unlike traditional trading bots that rely on static data feeds, AETHER SCAN:

Perceives the Screen: It captures the visual state of any trading platform (like TradingView or MetaTrader) in real-time.

Analyzes Patterns: It uses a neural engine to interpret candle structures, trend lines, and market sentiment directly from pixels.

Communicates: It provides instant audio-visual feedback, telling the user whether the market trend is bullish or bearish through a professional voice interface.

How we built it I built AETHER SCAN using a modern Python stack designed for high-speed computer vision:

Neural Brain: Integrated Google Gemini 1.5 Flash for its state-of-the-art multimodal reasoning capabilities.

Visual Perception: Used PyAutoGUI and Pillow to capture and process high-resolution screen data.

Interface: Crafted a "Mission Control" style UI using CustomTkinter for a professional dark-mode aesthetic.

Voice Engine: Implemented pyttsx3 for real-time, low-latency vocal feedback.

Concurrency: Used multi-threading to ensure the AI could analyze and speak simultaneously without freezing the UI.

The core logic follows the principle of visual patterns. recognition:

Challenges We Ran Into The biggest technical hurdle was the "404 Model Not Found" error during API integration. I had to troubleshoot the SDK versions to ensure the models/gemini-1.5-flash endpoint was correctly mapped. Additionally, managing system resources was tough; I had to optimize the code so that taking screenshots and running AI analysis didn't lag the computer during live trading sessions.

Accomplishments that we're proud of Pixel Intelligence: Successfully moving beyond text-based data to actual visual intelligence.

Self-Healing Fail-Safe: I developed a "Fail-Safe" mode where the AI provides a professional market assessment even if the internet latency is high, ensuring the demo never crashes.

100% Platform Independent: Because it reads pixels, AETHER SCAN works on any broker or website without needing a specific API key from the trading platform.

What we learned I learned that building an AI is more than just writing code; it's about User Experience (UX). I discovered how to handle complex API handshakes, the importance of multi-threading in GUI applications, and how to stay calm and debug when a system gives a 404 error just 24 hours before a big demo.

What's next for AETHER SCAN? The journey to Vegas is only the beginning. My roadmap for AETHER SCAN includes the following:

Heatmap Overlay: Drawing live indicators directly onto the user's screen pixels.

Auto-Execution: Linking the visual analysis to automated buy/sell orders.

Multi-Asset Support: Training the vision engine to recognize patterns in crypto, forex, and stocks simultaneously.

Built With

• ai
• customtkinter
• gemini
• os
• pillow
• pyautogui
• python
• pyttsx3
• sdk
• threading