How to Verify a Claim, Image, or Screenshot
About this check
If you have just seen a claim and want to know whether it is true, false, or not settled by the available evidence, start with three things: the exact wording, the original source, and the evidence behind it. A finished answer should not be just “yes” or “no.” It should give you a verdict, explain the level of confidence, and show any sources the check found.
The manual way to verify a claim
-
Capture the exact claim. Copy the wording if possible. Keep qualifiers such as “ reportedly,” “new,” or “in this country,” because small changes can alter what the sentence means.
-
Find the earliest source. Look for the original post, article, video, document, or statement rather than relying on reposts. Note who made the claim, when it appeared, and whether the surrounding context changes its meaning.
-
Search the wording. Put distinctive phrases in quotation marks, then try variations of the search. Look for reporting from independent sources, primary documents, official statements, or direct data.
-
Check the date. A claim can be accurate at one point and misleading later. Confirm that the sources you find refer to the same event, place, people, and time period.
-
Compare the evidence. Separate direct evidence from commentary. Check whether a source actually supports the claim or merely repeats it. If sources disagree, examine what each one can establish and what remains uncertain.
-
Write a qualified conclusion. Mark the claim as supported, contradicted, or unresolved. Keep the conclusion narrower than the evidence when necessary.
For a text-only claim, the manual process may take a few searches or a much deeper review. There is no single reliable time for every claim: a clear primary source may settle one quickly, while an old, edited, technical, or widely reposted claim may require more checking. The method also stops working when the original source is unavailable, the evidence is incomplete, or the available sources repeat one another without adding proof. In those cases, the responsible answer is uncertainty—not a forced yes or no.
Checking an image or screenshot
Images need a slightly different process. First, separate what the image shows from the caption attached to it. A real image can be paired with a false date, location, or explanation.
Look for the original upload and compare versions for cropping, edits, captions, and dates. Search distinctive text from the image. If it is a screenshot, read the text as a claim in its own right and check whether the alleged account, page, or document can be found elsewhere. If you cannot establish the source or context, label the image unverifiable rather than treating uncertainty as proof that it is fake.
A structured way to start with VOM
VOM accepts voice input for submitting a claim to verify, so you can state the claim instead of retyping it. For an uploaded image, VOM can assess whether it is likely authentic, likely AI-generated, or unverifiable. It can also read text inside a screenshot, allowing the forwarded image to be checked without retyping the text.
For a checked claim, VOM uses real-time web search and cites the sources behind its verdict. Its verification response includes a verdict, a confidence score, and any sources the check found. The verdict can be true, false, or inconclusive. When the available evidence cannot settle the claim, it reports inconclusive rather than forcing the result into yes or no.
VOM also supports natural back-and-forth conversations for everyday questions, learning, brainstorming, writing, and planning. That conversation experience is separate from the verification response: when you submit a claim for checking, the relevant result is the verdict, confidence score, and cited sources.
For people who want to follow circulating claims, VOM offers local, world, and trending feeds. It answers in the language being used across its eight supported languages. Together, these features give you a way to submit a claim or image, inspect the result, and check the cited evidence rather than relying on an unexplained label.