Source Credibility Checker
Paste a link or site address. This website credibility checker reads the address for fake-news red flags, then walks you through the questions a librarian would ask about the article itself.
Step 1 — The web address
Everything runs in your browser. The only thing that leaves your device is the bare site address, sent to the public registry lookup at rdap.org to find out when it was registered.
Read this before you trust the number. This is a media-literacy aid built on rules of thumb about web addresses, not a verdict on whether one article is true. A high score does not make a story accurate, and a low score does not make it false. Established outlets run corrections every week, and a brand new site can break a real story. Use the score to decide how hard to verify, never as the verification itself.
How to check if a source is credible
What this credibility checker looks at
Half the score comes from the address itself, which you can judge before reading a word: whether the link is encrypted, whether the address imitates a famous outlet, whether it sits on a cheap throwaway ending, and how long it has been registered. The other half comes from you, answering six questions about the article after you have read it. Neither half works alone. A clean address can host nonsense, and a scruffy one can host good reporting.
Lookalike addresses
Read a web address from the right. The part just before the ending is the part that is owned. In bbc.com.news-daily.net the owner is news-daily.net, not the BBC. Watch for a zero doing the job of an o, an l replaced by a 1, an extra hyphen, an extra word bolted on (cnn-news24.com), or the xn-- prefix, which means the address hides characters from another alphabet that look identical to Latin ones.
How old the address is
We ask rdap.org, the public registry lookup, when the address was first registered. A site that claims to be a long-running newspaper but registered its address three weeks ago has some explaining to do. Not every registry publishes the date, so a blank here is common and means nothing on its own.
The six questions, in plain words
Librarians teach a version of this as the CRAAP test, written at the Chico State library and used in classrooms ever since: currency, relevance, authority, accuracy, purpose. Stripped of the acronym it comes down to this. Is it recent enough for the claim to still hold? Does it answer the question you actually asked? Is a named person responsible for it, with an About page and a way to reach them? Can you follow the links and see the evidence yourself? And who benefits if you believe it, because a page selling you a supplement is not describing a study the way a researcher would.
That checklist has one weakness worth naming: it keeps you on the page, reading down. A site built to fool you can satisfy all five criteria beautifully, because it wrote the answers itself. That is why one of the six questions above sends you off the page entirely: is anyone else reporting this?
Read sideways, not down the page
Stanford researchers Sam Wineburg and Sarah McGrew watched professional fact-checkers, PhD historians and university students judge unfamiliar sites. The fact-checkers were quickest and most accurate, and they did something the others did not: they left the site almost immediately and searched for what the rest of the web said about it. Wineburg and McGrew named the habit lateral reading. Mike Caulfield turned it into four steps he calls SIFT — stop, investigate the source, find better coverage, trace the claim back to where it started.
In practice that is one extra browser tab. Search the site name with the word owner, funding or controversy. Search the headline in quotation marks and see who else carries it. If the claim rests on a study, a leaked file or a video, chase the original rather than the write-up. Do not settle for the About page: a site writes its own About page, so it tells you what the owners want you to think, not who they are.
Where the score is weakest
The known-outlet list holds around forty names, so most of the web comes back as unrecognised. That is not a mark against a site, it only means nobody vouched for it here. Satire sites are capped on purpose because a well-written joke reads exactly like a report. And no address check can tell you whether the numbers inside an article are right.
Bias and credibility are two different questions
A slanted paper can still get the facts right
People search for a bias and credibility checker as though those were one measurement. They are not. A newspaper can lean left or right and still report accurately, and a site with no visible politics can still print things that are simply wrong. The services that rate news keep the two apart, and it helps to know which one you are reading when you look a source up.
| Rater | What it reports | How it decides |
|---|---|---|
| Media Bias/Fact Check | A left-to-right placement plus a separate factual-reporting grade | Human reviewers, thousands of outlets, free to read |
| Ad Fontes Media | A chart with political lean on one axis and reliability on the other | Trained analysts from across the spectrum score sample articles |
| NewsGuard | A 0-100 trust score per publisher, shown as a browser label | Journalists grade nine published criteria and revisit ratings |
| This page | Address red flags plus the answers you give about the article | Rules of thumb, run in your browser, no rated-site database |
None of them, ours included, settles one article. A rating describes the outlet you are reading, which is a fair guide to how much checking a story deserves and no substitute for the checking.
Sites that are built to look local
The hardest sources to judge are not the shouty ones. They are the quiet sites with a town name and a tidy masthead. NewsGuard counted 1,265 of these partisan-backed sites dressed as local papers by June 2024, more than the 1,213 daily newspapers still operating in the United States, and by June 2026 its tracker listed 3,749 sites publishing machine-written articles across sixteen languages. Most are chasing ad money rather than votes, but both kinds read like news. When a local outlet you have never heard of turns up in your feed, the address age check and a search for the paper’s name do more work than reading the article again.
Checking a source for an essay or citation
How to check an author
An author credibility check is a search for the person, run away from the page they wrote. Look for a byline history: have they published elsewhere, and on this subject? Is there an institution, a job title or a professional body attached to the name, and does the institution list them? Do other people in the field cite or argue with them? A bare “staff writer”, a stock photo and a one-line biography are not proof of anything on their own, but on a page making a strong claim they are a gap worth closing. Pay attention when the subject sits outside the author’s training, which is where confident writing tends to outrun the evidence.
How to check a journal
Predatory journals charge an author a publication fee and skip the peer review the fee is meant to pay for. They are easy to mistake for real ones, so use the free checks that already exist rather than judging by the website:
- Think. Check. Submit. publishes a plain checklist for testing whether a journal or publisher is genuine.
- DOAJ, the Directory of Open Access Journals, lists open-access titles that have passed a vetting process, and removes ones that stop meeting it.
- COPE membership shows a publisher has signed up to published ethics standards.
- Retraction Watch tracks retracted papers and runs a hijacked-journal checker for clone sites that copy a real journal’s title and ISSN to collect submission fees.
- Scopus or Web of Science indexing tells you the title is inside the searchable record rather than outside it.
Check a paper for a retraction notice before you cite it. A retracted study keeps circulating in news write-ups and reading lists long after the journal has pulled it.
What your marker will accept
Cite the primary source, which is the study, the dataset, the court filing or the interview itself. A news article about a study is a secondary source: good for context, weak as the thing you hang a number on, because the number has already passed through a headline writer. Wikipedia is an excellent place to find sources and a poor place to cite one, so read its reference list and follow it out. Machine-written summaries are worse still, since a chatbot will happily produce a citation that does not exist. Whatever a general rule says, your course or journal guidance wins: some subjects accept preprints, newspapers or trade press, and others refuse all three.
Six mistakes people make judging a source
The tells that are not tells
Reading the padlock as a verdict.
HTTPS certificates are free and automatic. The padlock says your connection is private, not that the page is honest. Its absence is still a warning; its presence is not a clearance.
Believing the address ending proves ownership.
Anyone can buy a .org, and campaign groups and lobbying outfits do. Only a few endings are actually gated, such as .gov and .edu in the United States, and even those tell you who registered the address rather than whether one page on it is right.
Counting repeats as confirmation.
Twenty sites carrying the same story often means nineteen of them rewrote one wire report. Independent confirmation means a second outlet that did its own reporting, with its own named sources.
Taking the About page as evidence.
It is written by the people you are checking. Search the site name and read what others have published about its owners and funding instead.
Judging by design.
A clumsy template does not make a site wrong, and a polished one does not make it right. Content farms buy the same themes real publishers use.
Stopping at the headline.
Headlines are written to be shared, sometimes by someone who never read the piece. Check that the body actually supports the claim in the headline before you pass either on.
A short glossary
Words that come up when you check a source
| Term | What it means |
|---|---|
| Lateral reading | Leaving a page to find out what the rest of the web says about it, instead of reading further down the page itself. |
| Primary source | The study, dataset, filing or recording itself, rather than an article describing it. |
| Typosquatting | Registering an address one keystroke away from a real one, so a slip of the finger lands on your copy. |
| Punycode | The xn-- prefix that lets an address use non-Latin letters. Useful in most of the world, and abused to build addresses that look identical to famous ones. |
| Registration date | When the address was first bought. Published for most endings through the public registry, and a rough floor on how long the site can have existed. |
| Content farm | A site that publishes in bulk for search traffic and ad money, with little or no editing behind it. |
| Retraction | A journal withdrawing a paper it has published. The paper stays online, usually stamped, and should no longer be cited as evidence. |
Questions people ask
How do I check if a website is credible?
Start with the address: is it encrypted, does it imitate a bigger name, how old is it, and does the ending look like a bargain-bin ending. Then read the page and look for a named author, a date, working links to sources, and a real About page. Last, open a second tab and search the main claim to see whether anyone else is reporting it. This tool runs the first part for you and prompts you through the rest.
Is this a free website credibility checker, and does it work online without signing up?
Yes. It is free, there is no account and nothing to install. The checks run in your browser as soon as the page loads. The only network request is the registration lookup, and that sends nothing except the bare site address you pasted.
Can it tell me whether one article is true?
No, and be wary of anything that claims it can, chatbots included. The score describes the source and the shape of the article, not the facts inside it. A trusted newsroom can still get a story wrong, and an unknown blog can still be right. Treat the score as a signal for how much checking the claim deserves.
How is this different from the EasyBib website credibility checker?
EasyBib grades sites it has already reviewed and pairs that with its citation generator, so a site outside its list comes back with nothing to show. This page takes the opposite approach: it reads the address itself, which works for any site including one registered yesterday, then asks you the questions a librarian would ask about the article. If you need a citation at the end of it, you still need a citation tool.
How do I check whether an author is credible?
Search the name on its own, away from the page they wrote. Look for other work under the same byline, an institution or professional body that lists them, and whether people in the field cite or argue with them. Check that the subject matches their training. A story with no named author, or one credited only to staff, is not automatically wrong, but on a strong claim it is a gap worth closing before you rely on it.
What makes a source academically credible?
For an essay or citation the bar is higher than for reading the news. You want a named author with relevant expertise, a publisher or journal with editorial review, a date, and references you can follow. Peer-reviewed journals, government statistics agencies and university presses clear that bar. Most news sites, even good ones, count as secondary sources, and personal blogs and machine-written content usually do not count at all. Check your course guidance, since some subjects accept sources others reject.
How do I check if a journal is credible?
Use the free checks built for exactly this. Think. Check. Submit. publishes a checklist for testing a journal or publisher. DOAJ lists vetted open-access titles. COPE membership shows a publisher has signed up to ethics standards. Retraction Watch tracks retractions and runs a hijacked-journal checker for clone sites that copy a real title and ISSN. Indexing in Scopus or Web of Science is another positive sign. A fee to publish is normal in open access; a fee with no real peer review behind it is the predatory pattern.
How do I spot a fake news site?
The common tells cluster: an address that borrows a famous name, an ending like .xyz or .buzz, a registration date measured in weeks, no author, no date, no About page, a headline much louder than the body, and no other outlet carrying the story. One of those means little. Four or five together is your answer.
What about satire sites like The Onion?
They are recognised and flagged here, and the score is capped when one comes up. Satire is not lying, it is comedy, but screenshots of satire headlines circulate stripped of context all the time. If a story sounds too neat, check the front page of the site it came from.
Does an HTTPS padlock mean a site is trustworthy?
It means the connection is encrypted, nothing more. Certificates are free, so fraudulent sites have them too. Missing HTTPS is a genuine warning sign in 2026, but having it proves only that nobody is reading your traffic in transit.
Is this a bias and credibility checker?
It checks credibility signals, not political lean. The two questions are separate: an outlet can slant hard and still report accurately, and a site with no visible politics can still be wrong. For a lean rating, services such as Media Bias/Fact Check and Ad Fontes Media publish per-outlet placements alongside a factual-reporting grade, and NewsGuard scores publishers on transparency and accuracy criteria.
Does it check an Instagram or YouTube audience for fake followers?
No. An audience credibility checker is a different kind of tool: it samples the followers of an account to estimate how many look like bots or bought accounts, which matters if you are paying for a sponsored post. This page judges whether a website or article is worth trusting, so it has nothing to say about a creator's following.