How Our AI Sentiment Analysis Works

Every news headline on this page is scored by an AI language model trained to identify the directional bias of financial news. Each article receives one of three scores: Bullish (positive for the asset class), Bearish (negative), or Neutral (informational without clear directional bias).

The sentiment percentage shown above each news box is calculated from the aggregate of all scored articles in that category. A reading of 70% Bullish means that 70% of current news stories for that asset class carry a positive directional signal — and 30% are either bearish or neutral.

News is sourced in real time from Reuters, CNBC, Yahoo Finance, Forexlive, CoinDesk, Cointelegraph, and Finnhub. Sources are weighted equally. The sentiment model processes headline and summary text only — it does not analyse full article content.

News sentiment is one of the strongest short-term price drivers in liquid markets. Academic research consistently shows that sentiment shifts precede price movements by minutes to hours — particularly in forex and equity indices. Monitoring sentiment in real time gives traders an early read on directional pressure before it appears in price action.

High sentiment divergence — when news is heavily bullish but prices are falling, or vice versa — often signals elevated volatility ahead. Institutional risk managers use sentiment data to adjust position sizing and hedge ratios around major news events. A sudden shift in sentiment score from neutral to strongly bearish is a reliable signal to tighten stop-loss levels.

Sentiment is not a standalone trading signal — it is one input among many. Markets regularly move against prevailing sentiment, particularly around major policy announcements where the "buy the rumour, sell the news" dynamic applies. Sentiment is most useful as a confirmation tool alongside technical and fundamental analysis, not as a primary trigger.