How CryptoHypeBuster de-hypes crypto news
How aggregated crypto headlines are rewritten without hype, scored 0-100, and counted — and what the badges mean.
1. Aggregation
CryptoHypeBuster continuously aggregates headlines from 16+ crypto news sources via RSS, then deduplicates stories so the same wire story from multiple outlets appears once with all its sources listed.
2. Rewrite without the hype
Every deduplicated story is processed by a language model that does two things:
- produces a de-hyped rewrite that states the verifiable facts — who announced what, how much, when — and lists the exact hype phrases it removed;
- assigns a hype score from 0 to 100 measuring how much hype language flavored the original.
3. Scores and badges
- Hype meter — the 0-100 hype score shown as a 10-tick meter; 80+ renders in amber, below that in cyan.
- −N phrases pill — the number of hype phrases the rewrite cut from the story. Click through to the article page to see the original and the de-hyped rewrite side by side.
- ✓ clean pill — shown only when the story got a real de-hyped rewrite and the model cut zero phrases.
4. When processing fails, no badges are invented
Hype extraction is an LLM call and occasionally fails (provider rate limits, malformed output). In that case we persist a "Heuristic fallback" explanation with the story and the app suppresses all hype badges for it — a story is never marked CLEAN just because we could not analyze it.
5. What we count
Every phrase cut from a de-hyped rewrite is counted. The live stats page shows the running total, the number of stories de-hyped, and a day-by-day breakdown of the last 7 days.
Everything here is automated and can make mistakes; check the linked original sources before acting on any story. Not investment advice.