Check Facts
Start with the file. Upload the image, video or audio behind the claim and Verity tells you in plain English whether it looks captured, edited or AI generated — then work the five steps below to check the story around it. Free, unlimited, no account.
Checking a picture specifically? Use the image checker.
The five-step fact check
This is the order newsroom verification desks work in, compressed to something you can run on your phone in a few minutes. Doing them out of order wastes time: people research the story for an hour before noticing the photo is from another continent.
1. Write the claim down in one sentence
Most viral posts bundle several claims together: what happened, where, when, and who is to blame. Fact checking a blur is impossible. Strip it to a single testable sentence — 'this photo shows flooding in Valencia in October' — and check that one thing. If the sentence contains a word like 'destroyed', 'record' or 'banned', pin down what would count as true before you look for evidence.
2. Check the media itself before the story
The fastest failure mode for a viral post is a real caption over the wrong picture. Upload the file above: Verity reads the hidden camera data, the compression and resave history, and the grain and lighting patterns, and tells you whether it looks captured, edited, or generated. A file with no camera data and generator-like noise settles the question before you research a single word of the caption.
3. Find the earliest copy, not the loudest one
Run a reverse image search (Google Lens, TinEye, Bing Visual Search, Yandex) and sort by oldest. Recycled disaster and conflict footage almost always has an earlier posting under a different caption. For video, screenshot two or three distinct frames and search each — the thumbnail alone often fails to match.
4. Test the details against the physical world
Compare shopfront signage, language, license plates, road markings, vegetation, weather and the angle of shadows against maps and satellite imagery of the claimed location. Shadow direction and length pin down a rough time of day; a leafless tree in a claimed summer scene is enough on its own.
5. Look for independent confirmation, then say how sure you are
Two outlets both quoting the same anonymous post is one source, not two. Look for reporting with its own named witnesses, documents or on-site footage. When you pass the answer on, carry your confidence with it: 'the image is recycled from 2019' and 'I can't confirm this either way' are both useful; 'fake' with nothing behind it is not.
Sponsored
Six red flags worth a second look
None of these makes a post false. Each one raises the amount of evidence you should want before passing it on.
- The post tells you to spread it before it gets deleted.
- The account posted it for the first time and has no history in that subject.
- The image is a screenshot of a screenshot — recompressed until the evidence is gone.
- The emotional payload arrives before any verifiable detail: no place, no date, no name.
- The caption's numbers are round, extreme and unsourced.
- The only 'confirmations' are other posts quoting the same original post.
What a checker can and can't settle
It can settle: was this file captured?
Camera metadata, compression layering and sensor noise are physical residue of how a file came to exist. When they are intact, they distinguish a photograph from a render or a paste-up with real confidence, and they show which region of an image was altered.
It can't settle: is the caption true?
A completely genuine photograph can carry a completely false caption, and that is the most common form of misinformation by a wide margin. Nothing inside the file knows what someone typed underneath it. Steps three to five above are what answer that half.
It gets weaker with every repost
Platforms strip metadata and recompress everything on upload. A screenshot of a screenshot has almost no forensic evidence left, which is why "not enough to tell" is an honest answer rather than a failure — and why the original file is always worth chasing.
Common questions
- How do I fact check a photo or video for free?
- Upload it above for a free, unlimited authenticity read — no account needed — then run the five-step claim check on this page: isolate one sentence, verify the media, reverse image search for the earliest copy, test the visible details against maps and weather, and look for genuinely independent confirmation.
- What is fact checking, exactly?
- Fact checking is testing a specific, falsifiable claim against evidence that exists independently of the person making it. It is not deciding whether something sounds plausible or whether you trust the poster. A claim that cannot be stated precisely enough to be proven wrong cannot be fact checked at all.
- Can a fact check tool prove something is fake?
- No tool can. Authenticity analysis measures statistical traces — metadata, compression history, noise structure — and those traces degrade every time a file is screenshotted, cropped or reposted. Treat a detector result as one piece of evidence next to provenance, location checks and independent reporting, never as a verdict on its own.
- How do I fact check on Google?
- Use Google Lens or Images to reverse search the picture and sort results by date to find the earliest posting. Search the claim's distinctive phrase in quotation marks to see who else has used it. Google's Fact Check Explorer collects published checks from accredited organisations, and Street View plus satellite imagery let you match a claimed location against what is really there.
- Which fact checking organisations are worth trusting?
- Prefer organisations signed up to the International Fact-Checking Network's code of principles — they publish their sourcing and their corrections. Reuters Fact Check, AFP Fact Check, AP, Full Fact, Snopes and PolitiFact all show their evidence trail, which matters more than their verdict label: a check you cannot audit is just another claim.