AI-powered news collection, classification, and trading integration
The news intelligence system collects cryptocurrency news from multiple sources, classifies sentiment using AI, and integrates insights into trading strategies.
Each collected article is automatically classified by an LLM for sentiment and relevance:
Field
Values
Description
Sentiment
bullish, bearish, neutral
Overall market sentiment
Confidence
0.0 – 1.0
Model confidence in the classification
Relevance
high, medium, low
Relevance to active trading symbols
Symbols
List of tickers
Specific symbols mentioned in the article
Summary
Text
AI-generated one-sentence summary
{ "title": "Bitcoin ETF Inflows Hit Record $2B in Single Day", "source": "cryptopanic", "sentiment": "bullish", "confidence": 0.92, "relevance": "high", "symbols": ["BTC"], "summary": "Record institutional inflows into Bitcoin spot ETFs signal strong demand.", "published_at": "2025-07-15T10:00:00Z"}
News classification uses the system-wide LLM configured via HYPER_INSIGHT_SERVICE_TOKEN. For best results, use a model with strong reasoning capabilities.
News API providers typically have rate limits. Set fetch intervals that respect your plan’s limits. CryptoPanic free tier allows 5 requests per minute.