What we are trying to do
Any significant event is reported by dozens of outlets, and each one makes choices: what to lead with, whose account to carry, what to leave out. Read one and you get one shape of the story. That is not usually dishonesty - it is the ordinary result of finite space and a particular audience. But it means the picture you end up with depends heavily on where you happened to look.
Neutral is our attempt to make those choices visible. Every story we publish draws on several independent sources. We write it in plain language, set out the supportive and the critical readings on the same page, and attach a score describing how balanced the result is.
We are a small team doing this carefully rather than a large one doing it at scale, and we would rather say that plainly than imply more than we have built.
How the neutrality score works
The score describes one thing: how balanced a piece of writing is. It is not a verdict on whether the underlying events are good or bad, it is not a rating of the publications we drew on, and it is not a claim to have found the truth of a contested matter.
We assess each article against a consistent published standard. That standard asks whether the language carries emotional loading that the facts do not support, whether the reporting rests on a single account of events, whether opinion is presented as established fact, and whether a reader would be missing important context after finishing the piece.
The same standard is applied to everything we publish, which is the point of having one. Where a score would not be meaningful, we show none.
What the score bands mean
Multiple perspectives are represented, the language is measured, and the significant context is present.
The piece leans in places. That may be because the available reporting itself leans, or because one angle is better documented than another. Worth reading with that in mind.
The piece reads in a discernible direction. We publish it with the score attached rather than quietly adjusting it, so that you can weigh it yourself.
Where the machine ends and an editor begins
Over ninety per cent of our scores are produced automatically against the standard above. Consistency is something software does better than people, and applying the same test to every article is exactly the kind of work it should do.
A small number are adjusted by an editor. That happens when a piece passes the automated read but a person who has looked across the full coverage can see something the standard cannot.
There are things a checklist cannot see. A piece can be scrupulously even-handed in its language while drawing entirely on outlets that share a single institutional view - technically balanced, substantively not. Promotional material can be written in the register of reporting convincingly enough to pass an automated read. An editor who has looked across the full spread of coverage on an event can recognise both, where a rule applied to a single article cannot.
When an editor adjusts a score, that adjustment is recorded.
Corrections, and your say
We will get things wrong. When we do, we would rather hear about it and fix it, and we will say what changed.
This is an early version of something we intend to keep improving. The app lets you register your own read of any article, and we pay attention to what comes back - as we grow, that input is meant to shape how this methodology develops. We would rather build that carefully than claim it works better than it currently does.
These scores are our reading, published against a standard you can see, so that you have something specific to disagree with.
Write to us at contact@neutralperspectives.com.
Frequently Asked Questions
The ones we are asked most.
Is the app truly unbiased, and is the score ever wrong?
No, and we will not pretend otherwise.
This app is an honest attempt to bring balance into plain sight.
We calculate these scores primarily through linguistic evaluation driven by data. This involves analysing the structure and style of the text along with the facts present, supported by editorial review where it is needed.
Ultimately, we place the summary alongside supportive and critical perspectives directly in front of you. We let you read and decide for yourself, and then you can vote your opinion on the article you just read. By refining this systematic approach day by day, we aim to consistently deliver balanced global perspectives.
What does this tool actually look at?
The same standard we apply to our own writing, minus the one thing a single link cannot show. It reads the piece itself: whether the language carries emotional loading the facts do not support, whether it rests on a single account of events, whether opinion is presented as established fact, and whether important context is missing.
Does it look at where the story came from?
No, and that is the honest limit of scoring a single link. One article cannot show how widely an event was reported or whose account was left out, so we do not deduct for something it could never demonstrate. Judging breadth takes more than one piece, and that is where most of our own work happens.
Is this the same as the score on a Neutral article?
No, and the two are not comparable. This comes from a single document. A score on one of our own articles is produced across multiple sources with human review.
What if I paste a Neutral article?
You get that article's published score rather than a fresh one, so this tool and the article page never disagree with each other.
How many can I run?
Three a day. Every score costs a small amount to produce, so there is a daily cap that keeps the tool available to everyone. If you reach it, it resets the next day.
Why would a person change a score?
Because there are things a consistent standard cannot see on its own. An article can be scrupulously even in its language while drawing entirely on outlets that share a single view. Promotional writing can imitate reporting closely enough to read as news. An editor who has looked across the full spread of coverage can recognise both. When one adjusts a score, that adjustment is recorded.
How do I have my say on a story?
In the app. Every article carries a vote, so you can record your own read of it alongside ours. Neutral News is free on Google Play and the App Store.
Why does this app exist?
Because we were tired of reading multiple versions of the same event and still not knowing what happened. Most outlets hand you a verdict along with the facts. We would rather hand you the facts, show you the disagreement honestly, and let the verdict be yours.
Is it free?
Yes. No account, no paywall, no ads, and nothing to buy.
What makes this different from a news aggregator?
An aggregator collects headlines and leaves you to sort them out. We read across the coverage, write the story in plain language, put the supportive and the critical readings on the same page, and attach a score saying how balanced the result is.
Do you use AI to write the news?
We use software as a tool in the process, not as a replacement for judgement. It helps gather perspectives and keeps the same standard applied to everything. Every piece is reviewed by a person before it reaches you.
I think you got something wrong.
Tell us and we will look. When we get something wrong we would rather correct it openly and say what changed than quietly edit it away.
Can I tell you about a story you missed?
Please do. Write to us at contact@neutralperspectives.com. Small teams miss things and we would rather hear it from you.
Can I help?
Share it with someone who would find it useful. That is honestly how we grow, and it costs you nothing.