AI fact checker with sources
An AI fact checker is a tool that automatically verifies whether a claim is supported by independent sources. Justifact extracts the verifiable claims from a text, article, image or clip, searches the web in real time, compares the claims against the evidence found and publishes a verdict together with the list of sources consulted.
What makes an AI fact checker different
An AI fact checker does not answer from the model's memory; it searches for sources at the time of the question. The difference matters: a model asked directly will produce a plausible answer even about events after its training cut-off. Justifact runs the search before the verdict and shows what it found, so the reasoning can be checked by the reader.
What you can check
Text written directly, news articles submitted by link, images and video clips. For video, the platform transcribes the content, identifies the claims and checks each one separately. For images, authenticity (manipulation, AI generation) is reported separately from the factual verdict — they are two different questions.
What the verdicts mean
TRUE: supported by independent sources. MOSTLY TRUE: essentially correct, with imprecisions. FALSE: contradicted by sources. MISLEADING: literally correct but framed to suggest a wrong conclusion. INCONCLUSIVE: public evidence is insufficient. SATIRE: humour or parody, not a factual claim.
The limits of an automated verdict
The verdict is an assessment of the public evidence available at the time of checking, not judicial expertise. Human review applies above a risk threshold, not to every analysis. When evidence is insufficient, the platform answers INCONCLUSIVE rather than forcing a verdict — a design choice, not a technical limitation.
Frequently asked questions
What is an AI fact checker?
An AI fact checker is a tool that automatically verifies whether a claim is supported by independent sources, searching the web at the time of the question. Unlike an ordinary chatbot, it does not answer from the model's memory but from the sources it finds, which it also displays.
Can I check content that is not in English?
Yes. Justifact accepts text, articles submitted by link, images and video clips in any language, and searches for sources globally, so the evidence can be in any language too. On this site the verdict and the analysis are written in English. It has a web app, Chrome and Firefox extensions and an Android app.
How reliable is an AI-generated verdict?
An AI-generated verdict is as reliable as the sources it rests on — which is why Justifact displays them in full. The confidence score reflects how consistent the evidence found is. Analyses with a negative verdict and low confidence are held for human review.
The claims describing the standard professional scope of an electrician: renovating and bringing old electrica — ADEVĂRAT
The historical statement that 6 million Jews were killed during World War II. — ADEVĂRAT
An article reporting that the New York City Council voted by a 40-7 margin to ban horse-drawn carriages in Cen — APROAPE ADEVĂRAT
The claim that coffee 100% cures cancer. — FALS
The headline claim asserting that an AI became a government minister in Albania, featured under the segment 'L — APROAPE ADEVĂRAT