Editorial & data policy

This page describes exactly how the numbers and the words on CryptoNewsAlert are produced — so you can decide for yourself how much to trust them.

Sourcing rules

Every factual claim in our articles traces to a primary source: SEC and CFTC filings, official project announcements and blogs, GitHub releases, block explorers, court documents, and our own licensed price data. Headlines from other crypto publications serve only as signals that something happened — we never fetch, quote, or rewrite other outlets' article text, and our publishing software refuses to load their pages at the code level, so the rule cannot be quietly broken. Their reporting is their work; rewording it would add nothing for you.

Which coins get covered

Coverage is earned, not bought. A coin gets a page only if it passes four tests, in order: its name resolves to exactly one active CoinMarketCap listing (lookalike tokens are never added); it ranks in the top 500 with at least $1M of daily volume in our licensed data; there is enough verifiable substance — real use case, tokenomics, coin-specific price drivers — to fill a research page without padding; and any documented controversy can be stated factually on the page. A coin we would have to flatter, or whose problems we would have to hide, is rejected. Every verdict, pass or fail, is logged with its evidence, and most candidates fail.

Data provenance

All market data — prices, market caps, volumes, supply figures, and charts — comes from CoinMarketCap under a commercial license. An automated job refreshes quotes hourly and rebuilds the site; every page shows the "Last updated" timestamp of the snapshot it was built from. If an update fails its data-quality checks, the site keeps serving the last verified numbers rather than publishing suspect ones — a stale timestamp is the visible symptom. Charts draw only real fetched data points: a coin listed eight months ago gets an eight-month chart, never an extrapolated one. And we say "refreshed hourly," not "live," because that is what the data actually is — our freshness language is required to match the real refresh cadence everywhere on the site.

Historical figures such as all-time highs are recorded from public aggregator data when a page is created and reviewed periodically; they are not computed from our hourly feed.

AI-assisted drafting, human-gated publishing

We disclose this plainly: article drafts are assembled with the help of AI software that collects primary sources and proposes text. The machine's role ends there. Before a draft is even eligible for review, it must state in two sentences what it adds that isn't available elsewhere; drafts without a defensible answer are discarded. A machine checklist then verifies structure, sources, and banned-language rules (no predictions, no advice) — and a named person reviews every claim against the cited sources and personally decides to publish, revise, or reject. Nothing publishes automatically; rejection is a normal outcome. The reviewer's byline appears on the article and links to their author page.

Corrections

When we get something wrong, the article is corrected promptly and visibly: the corrected article carries a dated correction note stating what was wrong and what changed, and its "Updated" timestamp is set. We do not silently rewrite published claims. Errors in price data correct themselves at the next hourly refresh — the timestamp tells you which snapshot you are reading. To report an error, use the contact page; corrections reports are read first.

No advice, no predictions

We do not publish price predictions, targets, or investment advice, and our publishing checks reject drafts containing them. We do not accept payment for coverage, do not run sponsored content, and do not use affiliate links. What the site earns and what its editor holds are on the disclosures page.