Our Methodology
Transparency is the foundation of trust. Here is exactly how our AI generates cryptocurrency predictions, calculates risk scores, and tracks accuracy. No black boxes, no marketing fluff.
🔬 The 5-Step Process
Real-Time Data Collection
Every 30 minutes, our system fetches live market data from CoinMarketCap Pro API for 500+ cryptocurrencies. This includes current price, 24h volume, 7-day change, market capitalization, and circulating supply. Backup sources ensure 99.9% uptime even when primary API has issues.
Sentiment & Context Analysis
Beyond price data, we incorporate the Fear and Greed Index from Alternative.me, recent news sentiment, on-chain transaction metrics, and historical pattern correlations. This contextualizes price movements within the broader market psychology.
AI Prediction Generation
Data is processed by Google Gemini AI with carefully engineered prompts that focus on technical analysis, historical patterns, and multi-timeframe correlations. The AI generates a target price range (low and high), confidence score, and reasoning.
Risk Score Calculation
Each coin gets a 0-100 risk score based on: historical volatility (30-day std deviation), liquidity (24h volume relative to market cap), market cap tier (large/mid/small cap), and prediction confidence. Lower score = lower risk.
Continuous Validation
All predictions are tracked against actual market prices in real-time. Accuracy metrics are published publicly on our AI Accuracy page. We publish both wins and losses—full transparency.
📡 Our Data Sources
⚠️ Honest Limitations
We believe in radical honesty. Here is what our AI cannot predict:
- Black swan events: Major hacks, exchange failures, regulatory shocks.
- Manipulation: Whale activity, coordinated pump-and-dump schemes.
- Macro shifts: Sudden Fed policy changes, geopolitical events.
- Project-specific news: Team departures, security breaches, partnership announcements.
Past performance does not guarantee future results. Always do your own research.