Crypto News Sentiment Analyzer

Crypto News Sentiment Analyzer aggregates articles from CoinDesk, CoinTelegraph, The Block, Decrypt, and Blockworks and uses Claude AI to score each article's sentiment (bullish/bearish/neutral), extract mentioned tickers, and assign an impact score (1–10). BYO Anthropic API key. $0.005/article analyzed.

News sentiment is a key input for crypto trading signals: major outlet coverage with high impact scores often precedes 30-minute to 4-hour price moves. This actor structures that signal for automated ingestion by trading bots, morning briefing tools, and sentiment dashboards.

---

Why use Crypto News Sentiment Analyzer?

1. Trading signal pipeline — Filter for sentimentFilter: "bullish" and minImpactScore: 7 to get a real-time stream of high-conviction bullish signals from major crypto outlets.
2. Morning briefing automation — Run with lookbackHours: 12 overnight and get a structured summary of what moved sentiment while you slept — before markets open.
3. Ticker-specific monitoring — Set tickerFilter: ["SOL", "AVAX"] to track only articles mentioning your holdings. Avoid BTC/ETH noise.
4. Sentiment dashboard — Track the ratio of bullish vs. bearish articles over time. Rising bearish % ahead of price drops is a consistent leading signal.
5. AI trading agent integration — Returns structured JSON that slots directly into LLM-powered trading agents as a news context input.

---

How to use

1. Get a free Anthropic API key at console.anthropic.com (2 minutes).
2. Open the actor on Apify Store and click Try for free.
3. Paste your API key in anthropicApiKey.
4. Set lookbackHours, tickerFilter, and minImpactScore to focus on the signals that matter to you.
5. Click Start. Returns structured article records with sentiment, tickers, and impact scores.

---

Input

Field	Type	Default	Description
anthropicApiKey	string	required	Your Anthropic API key
sources	string[]	all	Sources: coindesk, cointelegraph, theblock, decrypt, blockworks
maxArticles	integer	50	Maximum articles to analyze
lookbackHours	integer	24	Hours of news history to fetch
tickerFilter	string[]	all	Filter by tickers (e.g. ["BTC","ETH","SOL"]) — empty = all
minImpactScore	integer	0	Minimum impact score 1–10 to include (0 = all)
sentimentFilter	string	all	bullish, bearish, neutral, or empty for all
model	string	claude-haiku-4-5	Claude model — haiku is fastest/cheapest, opus for best accuracy

---

Output

{
  "title": "Bitcoin ETF sees record $1B inflows",
  "url": "https://coindesk.com/...",
  "source": "coindesk",
  "published_at": "2026-08-15T10:30:00Z",
  "sentiment": "bullish",
  "impact_score": 8,
  "tickers": ["BTC"],
  "summary": "Spot Bitcoin ETF records largest single-day inflow since launch, signaling strong institutional demand.",
  "analyzed_at": "2026-08-15T11:00:00Z"
}

Output fields

Field	Type	Description
title	string	Article headline
source	string	Publication (coindesk, cointelegraph, theblock, decrypt, blockworks)
published_at	string	ISO 8601 article publish time
sentiment	string	bullish, bearish, or neutral
impact_score	integer	1–10 market significance (10 = major market event)
tickers	string[]	Cryptocurrencies mentioned (BTC, ETH, SOL, etc.)
summary	string	1-2 sentence AI-generated summary of key market implications

---

Cost estimation

Pay-Per-Event: $0.005 per actor start + $0.005 per article analyzed.

Use case	Articles	Estimated cost
24h news sweep (50 articles)	50	~$0.26/run
High-impact only (>7 score, ~10)	10	~$0.05/run
Hourly refresh (12 articles)	12	~$0.06/run
Daily morning briefing	30	~$0.16/run

Note: Claude API usage cost from your Anthropic account is additional (Haiku: ~$0.001/article).

---

FAQ

Why is an Anthropic API key required? The actor uses Claude to read each article and assess sentiment — this is more accurate than keyword matching alone, which struggles with sarcasm, context, and nuanced narratives. Using your own key gives you cost control and avoids shared rate limits.

Which model should I use? Haiku is fastest and cheapest (default). Use Sonnet or Opus for higher accuracy on complex articles or when impact score precision matters. Haiku is accurate enough for most trading signal use cases.

Does it work for non-crypto news? The actor scrapes crypto-specific publications. For general finance news, the ticker extraction and sentiment scoring still work, but coverage is limited to what these outlets publish.

How does impact score 1–10 work? Claude rates each article on market significance: 1–3 = routine commentary, 4–6 = notable (sector news, protocol updates), 7–8 = significant (major exchange listings, regulatory decisions), 9–10 = market-moving (ETF approvals, major hacks, macro events).

---

Related actors

• Crypto Social Sentiment Aggregator — Reddit + Fear & Greed sentiment (retail community signal to complement news analysis)
• Crypto Regulation Tracker — Regulatory news that often generates high-impact sentiment signals
• Crypto ETF Flow Tracker — ETF flow data to cross-validate bullish news sentiment with actual institutional buying

Feedback

If this actor is useful, a quick review helps other algorithmic traders find it: Leave a review on Apify Store