Selection
Which events become stories, which are followed for days, and which never reach the page.
Free · no account · no article limit
Look up a news outlet and see where named rating organisations place it. Open a story and see how the outlets that ran it split between left, centre and right, what each side emphasised, and where one side is not covering it at all.
8,437
news outlets on file
3,430
carrying a leaning published by a named rating organisation
7,937
we have seen cover a story compared here
2,291
coverage comparisons published so far
The short answer
Media bias is a recurring pattern in editorial choices that favours or disadvantages a perspective. It can affect what gets covered, what is omitted, who is quoted, how events are framed, and which words or images lead. Political leaning is only one form of it.
Which events become stories, which are followed for days, and which never reach the page.
Relevant facts, context or voices left out of an otherwise accurate account.
The angle, order and context that tell a reader what the event is mainly about.
Who gets quoted, whose expertise is accepted and whether opposing claims face the same scrutiny.
Loaded verbs, labels, headlines, photographs and captions that add approval, blame or drama.
How prominently and repeatedly a story is presented compared with other events of similar weight.
Bias and accuracy are different axes. A report can be factually correct and still omit relevant context; an outlet can lean left or right and publish careful reporting. That is why this site keeps political leaning separate from factual reliability, and both separate from the Trust score given to one individual story.
The fastest practical check is comparative: read the headline and evidence, note who is quoted, then look at how several outlets with different leanings reported the same event. The omissions become visible only when another account contains them.
Useful context: Pew Research Center on how people distinguish news from opinion · Reuters Journalistic Standards. Neither source supplies the outlet ratings used here; those are attributed to their individual rating organisations on each profile.
What the coverage actually looks like
Of the 2,209 stories where we have measured how coverage split across the spectrum, 1,769 had a side that barely covered it at all. The typical story was carried by 83.1 outlets. An outlet describing itself as unbiased is making a claim about itself; the spread is the evidence.
1,769 of 2,209
stories where one side barely covered it
Of the 2,209 published stories whose spread we have measured
83.1
outlets per story on average
183,614 outlet reports across those 2209 stories
1,027
outlets on the widest story so far
Every outlet we saw run it, rated or not
3,430
outlets carrying a published leaning
Of 8,437 publishers on file; the rest are marked unrated, never rounded to the centre
1,475 of those 1,769 have been watched to a close, which is when we are willing to call the gap a blindspot rather than a count. Coverage keeps arriving for hours after an event, and a side that has published nothing this morning may publish by tonight. The sample is small and recent — 2,291 stories published in total, of which 2,209 carry a measured spread — and everything above is a share of that rather than of the press as a whole.
Sharper still: on 153 of them no left-rated outlet ran the story at all, and on 195 none rated right did. Those are headcounts of the outlets carrying a published leaning rating, which is a smaller set than the full source list on each story — an unrated outlet is set aside, never counted as centre.
The widest field so far was Apple CEO John Ternus posts video hinting at iPhone 18 Pro camera feature, at 1,027 outlets. Every one of them is listed on that page with a link to its own report, so the figure can be counted rather than taken.
Coverage last measured 20 Sept 2026, 01:01
What this site does
The comparison is the product rather than a feature of it, which is why none of it sits behind a payment.
Each story carries an account of how each part of the spectrum told it, written here from the reporting rather than lifted out of it, with the outlets behind each account named and linked.
Not a selection. The full list, each entry with the leaning a rating organisation has published for it and a link to the outlet’s own report, so any part of the comparison can be checked against the source.
Where a story has been followed to the end and one side never ran it, that is reported as a finding. While it is still being followed, the page says that instead — an absence we have not finished measuring is not yet a fact about the press.
How the coverage split is worked out
The bar on every story is arithmetic over published ratings, not a judgement made here. It is worth knowing exactly which part is ours.
Every publisher we can see carrying the story goes into the list, wherever it sits and whether or not anybody has rated it.
AllSides, Ad Fontes Media and Media Bias/Fact Check publish ratings for news outlets. Each one appears on that outlet's page here against the organisation that made it, with a link to their entry. None of them is ours.
Far left, left and leans left become Left; leans right, right and far right become Right; centre stays centre. An outlet nobody has placed is never rounded into the centre — it is set aside and counted separately.
The percentages are shares of the outlets carrying a rating. The source list under them is every outlet that ran the story, rated or not, so the two figures do not match and each page says which is which.
Worked example — a story on the site now
Percentages are shares of the 25 outlets carrying a published leaning rating. 39 of the 64 outlets we know ran this story carry no rating and are not counted in them. Coverage measured .
And the three scores on the same story
The three scores
Every story carries three numbers, and each one is published with the sentence explaining where it landed and why.
0–100 · higher is better
How well corroborated and evidenced the original reporting is. It earns points for sources named on the record, documents quoted directly and independent outlets carrying the same thing; it loses them where the central claim rests on anonymous sourcing alone. A low score is not an accusation of dishonesty — it means there is less here that you or we can check.
0–100 · higher is better
The quality of the journalism, separately from how well sourced it is: context, a right of reply, fact kept clearly apart from comment, specific figures and dates. It loses points for a press release reproduced with nothing added, and for a headline the body does not support.
0–100 · lower is better
How far the presentation runs ahead of the substance. Measured, proportionate language scores low. Above 60 means heavy clickbait, manufactured outrage, or a headline the article itself cannot carry. It is the one score where the low number is the good outcome, and the page says so wherever it appears.
The full scoring rules, the models that run them, the prompt revision in force and the reason the previous revision was replaced are all published. Read the methodology.
Live from the site
2,291 stories have been published. These are the newest, with the coverage split shown on the card where it has been measured.
New laws create felonies for seizing ballots and interfering with mail-in votes.
Former Chilean president announced decision via social media after informal UN Security Council voting.
The company will receive R$1 per litre for 30 days, with potential for extension.
Rome ready to deploy up to four vessels to protect vital shipping lane, defence minister calls for more naval presence.
Millions were left without electricity after the entire grid disconnected on Friday.
Former president launches poll on social media for supporters to choose a new name for AI.
Check an outlet
Every publisher that appears in a coverage list here gets a page: the leaning each rating organisation published for it, each one shown against its author and linked, plus our own record of how often we have seen it cover a story. 3,430 of the 8,437 outlets on file carry a published leaning; the rest are marked unrated rather than assumed to be centre.
Our own count
Three organisations rate the same publishers, and holding their verdicts side by side is the thing a page carrying all three can do. We counted.
188 of 310
Of the publishers rated by two or more of AllSides, Ad Fontes Media and Media Bias/Fact Check, three in five get different verdicts. Where all three have rated the same outlet, only 44 of 133 are unanimous.
25 of 133
Asked for an unbiased news source most people name the Associated Press. Two of the three raters place it left of centre. Reuters is the one household name that clears it.
The honest comparison
Ground News is a real product, built by people who have been at this longer than we have, and this section is not going to pretend otherwise. It can be read without paying, and it sells subscription plans covering its fuller analysis. What those include and what they cost are theirs to set, so check their site rather than our summary of it.
| MediaBias News | Ground News | |
|---|---|---|
| What it is | One page per story showing how the left, the centre and the right covered it, what each side emphasised, and every outlet that ran it. | A news aggregator that merges reports on the same event so the coverage can be compared side by side, with bias, factuality and ownership data attached. |
| Where the political leaning ratings come from | AllSides, Ad Fontes Media and Media Bias/Fact Check. Each verdict is shown against the organisation that published it and linked to their entry. None of them is ours. | The same three organisations, per their own about page. Their ownership categories are their own work. |
| What it costs to read a full coverage comparison | Nothing. No paywall, no account, no article limit, and no part of the comparison reserved for a paid tier. | There is a free version, and there are paid subscription plans covering the fuller analysis. What each includes and what it costs are theirs to set — their pricing page is the only accurate source for it. |
| How much has been compared | 2,291 stories published, 2,209 of them with a measured coverage split, and 8,437 publishers on file. Counted from the database as this page was served. | Their about page states nearly 60,000 articles a day from over 50,000 sources. That is their figure, not a measurement of ours. |
Everything in the right-hand column is taken from Ground News’s own pages, read on 4 August 2026, and nothing has been put there that we could not confirm from them. Their prices are not reproduced because they are theirs to change and we would be misreporting them within a season; their own site is the only accurate source for the current terms.
What we can tell you more about is this site. Three things are true here, and they are the reasons to use it.
No paywall, no account, no article limit, and no part of the comparison reserved for a paid tier. The intention is to pay for it with advertising and editorial work rather than with readers; nothing is sold on it yet, which is why there is no advertising on this page.
The scoring rules, the prompt revision that produced the scores you see, the reason the previous revision was replaced, and the plagiarism gate that rejects a draft reusing a source’s sentences are all on the methodology page. So is the corrections log.
We do not republish anybody’s paragraphs, and we do not re-badge anybody’s rating as an assessment of ours. Third-party ratings appear with the rater named. Our own writing is written from the facts, and every report it was written from is linked.
What is missing here is scale. This site started publishing in 2026 and has 2,291 comparisons up at the time you are reading this. Every number on this page is the real one, counted on the request, which is the only claim about ourselves we are in a position to make.
Questions
Media bias is a recurring pattern in editorial choices that favours or disadvantages a perspective. It can appear in which events are selected, which facts or voices are omitted, how a headline frames the event, which sources are trusted, the words and images used, and how prominently the story is placed. Political bias is one form; commercial, cultural, geographic and institutional pressures can shape coverage too. Bias is not the same thing as a factual error, and an outlet-level rating is not a verdict on every article.
Every newsroom makes selections, so no press system is free of perspective. That does not mean every choice is unfair or every report is false. A free press contains many independent newsrooms whose choices can be compared; the practical safeguard is plural coverage, transparent sourcing, corrections and a clear separation between reporting and opinion.
There is no defensible universal winner. Bias raters use different methods and often disagree, while outlets change over time and news, analysis and opinion can differ inside one brand. A useful comparison names the rater, the date, the sample and the dimension being measured instead of publishing an absolute league table.
Part of the change is the information environment: audiences are fragmented, opinion and reporting sit beside each other, and digital distribution rewards material that wins attention quickly. Part is perception. People judge the same coverage differently depending on their prior views, and trust in specific outlets is sharply divided by politics. That does not prove every newsroom has become more biased; it is why claims about a trend need content evidence, audience evidence and dates kept separate.
Ground News can be read without paying, and it also sells subscription plans covering its fuller analysis. What each plan includes and what it costs are theirs to set and change, so check their site for the current terms rather than taking ours. MediaBias News is free in full: no paywall, no account, no article limit, and no part of the coverage comparison held back for a paid tier.
We are not in a position to rank the field, so here is what this one is. MediaBias News publishes the same kind of comparison — which outlets ran a story, where each of them sits on the political spectrum, and what each side emphasised — without a subscription, an account or an article limit. It is younger and much smaller: 2,291 comparisons published and 8,437 outlets on file so far. Everything it holds is public.
Two checks, and they answer different questions. For the outlet, look up whether a named rating organisation has placed it — AllSides, Ad Fontes Media and Media Bias/Fact Check publish leaning and factual-reliability verdicts, and every outlet page here carries theirs with the rater named and linked. For a particular story, that is not enough: look at who else ran it and where those outlets sit, because an outlet's average leaning tells you little about one report. Each story page here lists every outlet that ran it and how they split.
Because what an outlet leaves out does not appear in what it publishes, and you cannot detect an omission from inside a single account. Of the 2,209 stories here where the spread has been measured, 1,769 had a side of the spectrum that barely covered them at all — a reader following only that side would not have known the story existed. The average story was carried by 83.1 outlets.
A story that one part of the political spectrum has effectively not reported. It is a claim about absence, so it is only made here once a story has been watched to the end and the side still ran nothing — coverage keeps arriving for hours, and a gap measured at nine in the morning may be filled by the evening. While a story is still being checked the same counts are shown with a note that the watch is open, which is a count rather than a finding.
A way of finding out where a news outlet sits on the political spectrum, and how a story's coverage was distributed across it. On this site it is two things. Each publisher has a page carrying the leaning that AllSides, Ad Fontes Media and Media Bias/Fact Check have published for it, with every rating shown against the organisation that made it. Each story has a bar showing how the outlets that ran it split between left, centre and right.
No single outlet is unbiased, and one advertising itself as unbiased is making a claim it cannot support. The nearest thing to unbiased news is the same event read as several outlets across the spectrum told it, with attention to what only some of them said. This site does not offer you a neutral outlet. It shows you the spread and links to every report in it, so the judgement stays yours.
There is no MediaBias News app in the App Store or on Google Play. The site installs to a phone or desktop home screen as a web app from the browser’s own menu, and pages already opened stay readable if the connection drops. Nothing has to be installed to read it.
We do not decide it. Each outlet carries the leaning that named rating organisations — AllSides, Ad Fontes Media and Media Bias/Fact Check — have published about it, shown on the outlet's page against the organisation that made it and linked to their entry. Where nobody has placed an outlet it is marked unrated and left out of the percentages rather than rounded into the centre.
Yes. Every story, every coverage breakdown and every outlet page is free, with no registration and no article limit. The intention is to pay for the site with advertising and editorial work rather than with subscriptions; nothing is sold on it yet, which is why there are no ad slots on this page.
It publishes what each side reported and links to it, so the comparison can be checked. Where outlets disagree on a fact, the disagreement is reported rather than settled by us. Where only one part of the spectrum has covered something, that absence is reported too — once the story has been watched long enough for the absence to mean anything.
Every weekday morning
The story the press told most differently today, and how far apart they were. Free, and one email — no upsell to a paid tier, because there is not one.