AI & disinformation detector
Real or fake?
Check in seconds.
Upload a photo, paste a link or text. Czytofejk will assess whether the content was AI-made and whether it may mislead — and show you the sources.
Upload a photo or video. We'll check whether it's real or AI-generated.
↑
Drag, click or paste a photo or video
PNG · JPG · MP4 · MOV · max 50 MB
or paste a link
Also works with:TikTokYouTubeFacebookInstagramX
The result is a probability assessment, not a verdict. We don't store your files.
How it works
From content to verdict in three steps
Full process description →01
Upload or paste
A photo, a link or text — anything that raises your doubts.
02
Signal analysis
We check AI-generator artifacts, metadata and manipulation patterns.
03
A clear verdict
A confidence percentage, a list of signals and the sources that back it up.
4
content types in one detector: images, video, social media clips and text
100%
of verdicts come with signals, quotes from the material and sources you can verify yourself
<1 min
first check — free, no account, no install
Radar · live
What Poles are checking right now
Full disinformation Radar →What we detect
Four content types, one detector
See how it works →AI images and graphics
An ensemble of independent detectors, generator artifacts, camera metadata and C2PA provenance.
Deepfakes in video
Frame-by-frame analysis of video files — synthetic faces and AI-generated footage.
Social media videos
Paste a TikTok, YouTube or Facebook link — we transcribe the audio, analyse the frames and the post description.
Fake news in text
Claim verification on the open web, matching against known disinformation narratives and fact-checker debunks.
Who it's for
One detector, four everyday use cases
Newsrooms and fact-checkers
Evidence, not just a verdict
Full video transcript, highlighted claims with a verdict for each, sources and a citable PDF report with the check date. Plus an API and team accounts with per-member limits.
Let's talk about a rollout →Online buyers and sellers
Listing photo looks too good?
Check whether a product photo from Vinted or a marketplace was AI-generated, and whether a suspicious link leads to a domain on the CERT Polska warning list.
Check a photo now →Conscious readers
Verify before you share
A sensational post, a chain message from family or a TikTok video: paste it and get a verdict with manipulation signals and sources you can inspect yourself.
Verify content →Public institutions
Disinformation narrative monitoring
The Radar shows live what Poles are checking, and analyses match content against the EUvsDisinfo narrative index and Polish debunks. API integration, teams with an operations audit trail.
Get in touch →Technology and scale
Built on our own technology, ready to integrate
We're not a wrapper around a single API. Our ensemble combines external detectors with our own models and narrative index — and you pay exactly the real cost of each analysis.
Own models
Our own GPU image and video detector, C2PA/EXIF forensics and a calibrated vote ensemble — no single-vendor dependency.
Narrative index
Thousands of cases from EUvsDisinfo and Polish debunks in our own vector database — new content is matched against known disinformation campaigns.
Public API
The same pipelines over REST: images, text and social media videos, sync and async modes, verdicts in 3 languages.
Fair billing
Credits reflect the real cost of each operation — we hold a deposit and refund the difference after the analysis. No subscription traps.
84
items checked in the last 30 days
63%
of them flagged as risky
19.7 s
median detector response time
Got questionable content? Check it now.
Your first check takes less than a minute. No account, no install.
Go to the detector →