Documentation

Comprehensive guide to understanding how the AI Market Analyser generates intelligent trading recommendations.

Table of Contents

System Overview

The AI Market Analyser is an advanced trading recommendation system that combines multi-timeframe technical analysis, AI-powered sentiment analysis, and intelligent decision-making algorithms to provide actionable trading recommendations for stocks, cryptocurrencies, and forex pairs.

Core Philosophy: Traditional technical analysis relies on single timeframes and manual interpretation. Our system analyzes multiple timeframes simultaneously, incorporates real-time news sentiment, and uses AI to synthesize complex signals into clear, actionable recommendations.

System Capabilities

Multi-Timeframe

1-min, 15-min, 1-hour analysis

News Analysis

Real-time sentiment from 50+ sources

AI-Powered

GPT-4 sentiment interpretation

184,799+ Symbols

Stocks, Crypto, Forex, ETFs

How It Works

When you analyze a symbol, our system executes a sophisticated 5-step process that typically completes in 10-20 seconds:

1

Market Data Collection

The system fetches historical price data for three timeframes (1-minute, 15-minute, and 1-hour) from the 12 Data API. This provides both short-term and longer-term market context.

Real-time Data Multi-Timeframe
2

Technical Indicator Calculation

For each timeframe, the system calculates multiple technical indicators including RSI, MACD, support/resistance levels, volume trends, and price momentum. This provides a comprehensive technical picture across different time horizons.

RSI MACD Support/Resistance Volume Analysis
3

News & Events Gathering

The system retrieves recent news articles (last 24 hours) related to the symbol from NewsAPI, which aggregates content from over 50 news sources including financial news outlets, mainstream media, and industry publications.

50+ Sources Real-time News
4

AI Sentiment Analysis

News articles are analyzed by OpenAI's GPT-4o-mini model, which interprets the sentiment, context, and potential market impact. The AI provides a sentiment score (-1.0 to +1.0) and explains its reasoning in plain English.

GPT-4o-mini Context-Aware Explainable AI
5

Recommendation Generation

All signals are combined using a sophisticated weighting algorithm. Longer timeframes receive higher weight (1-hour > 15-min > 1-min), and sentiment contributes 30% to the final score. The system then generates a clear BUY/SELL/HOLD recommendation with entry price, target price, stop loss, and risk assessment.

Weighted Signals Risk Management

Multi-Timeframe Technical Analysis

Why Multiple Timeframes Matter

Traditional technical analysis often focuses on a single timeframe, which can lead to contradictory signals or missing the bigger picture. Our multi-timeframe approach provides:

  • 1-minute timeframe: Captures immediate price action and short-term momentum (ideal for day traders)
  • 15-minute timeframe: Shows intraday trends and patterns (ideal for swing traders)
  • 1-hour timeframe: Reveals longer-term trends and market structure (ideal for position traders)

Technical Indicators Used

RSI (14)

Identifies overbought/oversold conditions

MACD

Detects trend changes and momentum

Support/Resistance

Key price levels for entries/exits

Volume Analysis

Confirms trend strength

Price Trend

Direction and momentum analysis

Volatility

Risk assessment and sizing

Signal Weighting Strategy

Not all timeframes are equal. Our algorithm uses intelligent weighting:

  • 1-hour signals: Weight = 3 (highest priority - stronger trends)
  • 15-minute signals: Weight = 2 (medium priority - intraday confirmation)
  • 1-minute signals: Weight = 1 (lowest priority - noise filtering)

This ensures that longer-term trends aren't overshadowed by short-term noise, while still capturing immediate opportunities when all timeframes align.

AI-Powered Sentiment Analysis

Beyond Traditional News Scanning

Traditional sentiment analysis tools simply count positive/negative words. Our AI-powered approach understands context, sarcasm, market impact, and relative importance of news events.

How GPT-4o-mini Analyzes Sentiment

The AI model receives recent news articles and performs sophisticated analysis:

  • Context Understanding: Differentiates between "earnings beat expectations" vs "earnings disappoint despite growth"
  • Impact Assessment: Evaluates whether news is material (e.g., CEO resignation vs routine product update)
  • Market Psychology: Considers how traders are likely to react to specific types of news
  • Temporal Relevance: Weights more recent news higher than older articles
  • Source Credibility: Considers the reliability and influence of news sources

Sentiment Output

The AI provides three key outputs:

  • Category: Positive, Negative, or Neutral (clear classification)
  • Score: -1.0 to +1.0 (quantitative measure for algorithmic use)
  • Rationale: Plain-English explanation of why the sentiment was determined (transparency & explainability)

Example: For a tech stock, "Company announces layoffs of 10% workforce" might be interpreted as slightly positive (cost-cutting measures to improve profitability) rather than negative, depending on the broader financial context. Traditional keyword-based tools would miss this nuance.

Recommendation Engine

Signal Synthesis Algorithm

The recommendation engine combines technical and sentiment signals using a weighted scoring system:

1

Technical Score Calculation

Each timeframe contributes a score based on its indicators:
Technical Score = (1-hour × 3) + (15-min × 2) + (1-min × 1)

2

Sentiment Integration

Sentiment contributes 30% to the final score:
Combined Score = Technical Score × 0.7 + Sentiment Score × 0.3

3

Action Determination

Based on the combined score, the system determines the recommended action:

  • Strong BUY: Combined Score > 3.0 (high confidence upward move)
  • BUY: Combined Score > 1.5 (moderate upward bias)
  • HOLD: Combined Score between -1.5 and 1.5 (neutral/uncertain)
  • SELL: Combined Score < -1.5 (moderate downward bias)
  • Strong SELL: Combined Score < -3.0 (high confidence downward move)

Risk Management Features

Every recommendation includes built-in risk management:

  • Entry Price: Current market price or optimal entry level
  • Target Price: Expected profit target based on historical volatility and trend strength
  • Stop Loss: Maximum acceptable loss level (typically 2-5% below entry)
  • Risk Level: Low/Medium/High classification based on volatility and signal strength
  • Confidence Score: Percentage indicating how strongly signals align (0-100%)

Key Advantages

Multi-Timeframe Perspective

Traditional tools force you to manually switch between timeframes. Our system analyzes all three simultaneously, catching opportunities and avoiding false signals that only become apparent when viewing multiple time horizons.

AI-Powered Context Understanding

Unlike sentiment tools that rely on keyword matching, our GPT-4 integration actually understands news context, market psychology, and relative importance. It can differentiate between "earnings beat by 2%" (positive) and "earnings beat by 2% but guidance lowered" (actually negative).

Speed & Efficiency

Manually performing this analysis (checking 3 timeframes, reading 10+ news articles, calculating indicators, synthesizing signals) would take 30-60 minutes per symbol. Our system does it in 10-20 seconds.

Removes Emotional Bias

Human traders often suffer from confirmation bias, FOMO, or panic selling. Our algorithmic approach makes objective, data-driven decisions without emotional interference.

Backtested Indicators

We use proven technical indicators (RSI, MACD, support/resistance) that have decades of academic research and real-world validation, not experimental or proprietary "black box" indicators.

Explainable Recommendations

Every recommendation comes with a detailed rationale explaining why the system made that decision. You're never left wondering "why did it say BUY?" - the reasoning is transparent and educational.

Smart Caching

Our caching system reduces API costs and improves speed. If someone analyzed AAPL 30 seconds ago, your analysis uses the cached data (it's still fresh!) without consuming your credits.

Cost-Effective

Professional trading tools with similar capabilities cost $50-200/month. We offer a credit-based system where you only pay for what you use, and cached analyses are completely free.

Comparison with Traditional Methods

Feature AI Market Analyser Manual Technical Analysis Basic Sentiment Tools
Multi-Timeframe Analysis ✅ Automatic (3 timeframes) ❌ Manual switching required ❌ Not available
Sentiment Analysis ✅ AI-powered context understanding ❌ Manual news reading ⚠️ Keyword-based only
Speed ✅ 10-20 seconds ❌ 30-60 minutes ⚠️ 5-10 minutes
Emotional Bias ✅ Objective algorithms ❌ Subject to human bias ⚠️ Mixed (depends on user)
Risk Management ✅ Automatic (entry, target, stop loss) ⚠️ User must calculate ❌ Not included
Explanation of Signals ✅ Detailed rationale provided ✅ User understands their logic ❌ Black box scores
Cost ✅ Pay-per-use (credits) ✅ Free (but time-consuming) ❌ $50-200/month subscription
Consistency ✅ Same logic every time ❌ Varies by trader skill/mood ✅ Consistent but limited
Scalability ✅ Analyze dozens of symbols quickly ❌ Limited to few symbols daily ⚠️ Can scan many but shallow analysis

Use Cases & Best Practices

Who Should Use This Tool?

  • Day Traders: Quickly identify intraday opportunities across multiple symbols
  • Swing Traders: Spot multi-day trends supported by both technicals and sentiment
  • Position Traders: Validate longer-term positions with comprehensive analysis
  • Crypto Traders: Navigate volatile crypto markets with objective signals
  • Forex Traders: Analyze currency pairs with technical + geopolitical news sentiment
  • Learning Traders: Study how professional-grade analysis combines multiple factors

Best Practices

Important: This tool provides recommendations, not guarantees. Always use proper risk management, never risk more than you can afford to lose, and consider recommendations as part of your broader trading strategy.

  • Cross-Verify Signals: Don't trade on a single analysis - use multiple symbols and timeframes
  • Respect Risk Levels: High-risk recommendations should use smaller position sizes
  • Use Stop Losses: Always set the recommended stop loss to protect your capital
  • Monitor Sentiment Changes: News sentiment can shift rapidly - re-analyze before major moves
  • Combine with Your Strategy: Use our recommendations as one input in your trading plan
  • Paper Trade First: Test the system's recommendations with virtual money before risking real capital
  • Track Performance: Use the History feature to review past recommendations and learn from outcomes

When NOT to Rely Solely on This Tool

  • During major market-moving events (Fed announcements, earnings releases, geopolitical crises)
  • For penny stocks or extremely low-volume instruments (data quality may be poor)
  • When your personal risk tolerance differs significantly from the recommendation
  • If you have insider knowledge or company-specific information the algorithm can't access

Ready to Get Started?

Start analyzing stocks, crypto, and forex with AI-powered insights.

Login to Start