How we know what we know.

The standard

Breaking news moves faster than certainty. Most of what goes wrong in live coverage comes from blurring that line, so we hold every update to one rule:

Each label carries a specific meaning:

• Developing Reported by credible sources but not independently confirmed. We tell you who is saying it and what is still unknown.
• Confirmed Traced to an official source, primary documentation, or at least two independent sources. We link the evidence.
• Corrected An earlier update was wrong. The correction states what we said, what was inaccurate, and what we know now.

The pipeline

Every story on Noteworthy News moves through the same five stages, whether it is a one-line wire update or a live story running for days.

1. 01 Detect

   Signals come from official feeds we monitor directly - USGS, the National Weather Service, government and agency alerts - plus primary accounts and local reporting we have vetted in advance.

2. 02 Verify

   Before a claim is published, it must trace to its original source. We check who said it first, whether they are in a position to know, and whether independent sources agree.

3. 03 Context

   We state plainly what is known, what is not, and who is doing the saying. If a number is preliminary or an account is disputed, the update says so in the same sentence.

4. 04 Publish

   Every update ships with a timestamp and a source note attached. Live stories keep a visible timeline, so you can see when each fact entered the record.

5. 05 Correct

   If we got it wrong, the correction is published in the same place readers saw the error. Nothing is silently rewritten after the fact.

Sources we accept

Not all sources carry equal weight. In order of preference, we rely on:

1. Primary documentation. Official statements, court filings, agency data, and direct records of the event itself.
2. Official channels. Verified government, emergency-service, and institutional accounts - checked against the organization's own site, not just a platform badge.
3. Direct witnesses. People verifiably present at the scene, weighed against physical evidence and other accounts.
4. Established reporting. Newsrooms with a track record on the specific beat, credited by name when we rely on their work.

A single anonymous claim is never enough. When a story rests on reporting we cannot independently confirm, we attribute it explicitly - "according to" is a load-bearing phrase in our updates, not filler.

Verifying images and video

Recycled and manipulated footage is the most common failure in breaking coverage. Before any image or clip appears in our reporting, it goes through four checks:

• Reverse image search. Has this exact frame appeared online before, attached to a different event?
• Geolocation. Do landmarks, signage, terrain, and architecture match the claimed location?
• Time consistency. Do weather, shadows, and light match the claimed date and hour? Does available metadata agree?
• Provenance. Who posted it first, and were they plausibly in a position to capture it?

Media that fails any check is either excluded or published only with an explicit warning about what we could not establish. When a viral clip is fake, we say so directly rather than leaving it unaddressed.

When we can't verify

Sometimes a claim is everywhere and we still cannot stand it up. In that case we have three options, and publishing it as fact is not one of them:

• Hold. If a claim is not yet load-bearing for the story, we keep working it and publish nothing.
• Publish with attribution. If a claim is significant and comes from a credible source, we publish it as developing, name the source, and state what remains unconfirmed.
• Address it head-on. If a false or unproven claim is spreading widely, we cover the claim itself - what is being said, where it started, and why it does not hold up.

Facts and analysis stay separated. When an update includes interpretation - why an event matters, what may happen next - it is framed as such, never mixed into the factual record.

Corrections in public

We will get things wrong. Live coverage guarantees it. What we control is what happens next:

• Corrections are posted in the same timeline where the error appeared, clearly labeled and timestamped.
• The original claim is not deleted or quietly edited; readers can see what changed and when.
• Material corrections are noted at the top of the affected story, not buried at the bottom.
• If an alert went out with an error, a corrected alert follows through the same channel.

Holding us to it

This standard only means something if readers can enforce it. If an update is wrong, missing a source note, or blends developing claims into confirmed fact, we want to hear about it - and we respond to correction requests before anything else in the inbox.

The full rules our newsroom operates under, including independence and funding, are in our editorial policy. The techniques we use are public too: fact-checking tips covers the exact checks described above, so you can run them yourself.