Is this image or video AI-generated? Check in 3 seconds

Detect AI-generated photos and videos for free. Check dating profiles, news, and social media images for AI generation probability and deepfakes instantly in your browser.

Detect AI-generated photos and videos for free. Check dating profiles, news, and social media images for AI generation probability and deepfakes instantly in your browser. Free with no sign-up, and the analysis runs inside your browser — files never leave your device. On our public gate (30 real + 52 AI images) it measured 96.3% at original quality and 91.5% on recompressed copies, with zero real photos confidently mislabelled as AI. Treat it as an indicator, not proof: the newest commercial generators evade every detector.

What this tool does better

• No upload — your photo never leaves the device
• Two-model ensemble cross-checked with C2PA provenance
• Published numbers: 94.7% pristine, 87.6% recompressed, 0 false flags

• AI image & video detector — Upload a photo or video to get an AI-probability score.
• AI Detection Accuracy — Public Report — Measured accuracy, comparisons, honest limits.
• AI Spotter Quiz — Daily 10 questions to train your eye.
• Was this written by AI? — Estimate AI authorship from style and repetition patterns.
• Is this review real? — Spot exaggeration and fake-review patterns.
• Human or bot? — Estimate bot likelihood from username, bio and posts.
• Is this link safe? — Find phishing signals from the URL structure alone.
• Is this message a scam? — Combine link risk with bait phrasing to flag smishing.
• Is this QR safe? — Decode a QR image and risk-check the link inside.
• Is this email really from them? — Catch sender spoofing from domain and body.
• Giveaway DM Check — Paste a "you won!" DM and find the scam patterns.
• Impostor Handle Check — Catch impostor handles using look-alike letters.
• Ticket Resale Scam Check — Check resale-ticket red flags before you pay.
• Marketplace Listing Check — Screen listings and sellers for scam signals.
• Romance Scam & Catfish Check — Check whether a match is a real person.
• Fake Screenshot Check — Check whether a circulating tweet/chat capture is forged.
• Reverse Image Search Guide — How to trace a photo’s origin with reverse image search.
• Platform AI Labels Guide — What platform AI labels mean — and don’t.
• Call Recording Scam Check — Transcribe a call and scan for scam phrases.
• AI Disclosure Self-Check — Self-check your AI disclosure duties.
• Original Registration Notary — Register a file hash to prove "it existed at this time".
• Provenance · Content Credentials — Read C2PA/EXIF data to identify the generator.
• Is your photo leaking data? — Check and strip location data in photos.
• Trends — See the verdict mix across your own check history.
• Install & integrate — Install to your home screen — works offline too.

How it works

1. Drop a file

Just drag in a photo or video. No sign-up, no upload.

2. Analyzed on your device

Detection models run in your browser for instant results. Analysis data helps improve detection quality.

3. See & share

Get an AI-probability gauge and share the result link with a tap.

Detection algorithm & accuracy

Two benchmark-proven detectors — SwinV2 and Community Forensics (CVPR 2025, trained on 4,800+ generators) — run as an ensemble. Their failure modes differ (photorealistic people vs. AI art), so together they are far more robust. On top of that we check file provenance metadata (C2PA/IPTC AI markers): generators like Firefly, DALL-E and Gemini leave official markers that we catch definitively, independent of pixels. For video we sample multiple frames and aggregate over time.

Internal benchmark (40 real + 70 AI incl. DALL-E 3 & Gemini, with social-media recompression): ensemble accuracy 98.2% · 29/30 Gemini detected · zero real photos falsely confirmed as AI. Detector v2.1 retrained in-house on latest-generator data.

Note: the top 2026 commercial generators (GPT Image 2, Nano Banana Pro, MAI-Image-2.5) are hard for any detector (18–30% accuracy in studies) — photorealism has passed what the eye alone can settle. Compressed, screenshotted, or upscaled images also lower accuracy.

• Two-model ensemble (SwinV2 + Community Forensics) — selected via hands-on benchmark
• Provenance check (C2PA/IPTC/generator signatures) → catches the newest commercial generators
• Video: frame sampling + temporal aggregation (mean + peak)
• Runs fully in the browser → privacy + zero server cost

FAQ

Do you upload my photos?

Analysis runs in your browser. To improve detection quality, analysis data may be stored — see our privacy policy for details.

Is it really free?

Yes — free, no sign-up, no limits. The service is supported by ads.

Can I fully trust the result?

No. It is an indicative signal. Newest generators and post-processed images can cause errors.

Which files are supported?

JPG, PNG, WEBP, GIF images and MP4, WEBM videos.

Can I check if a dating-app match is a real person?

Save their profile photo and drop it in. A benchmark-selected SwinV2 detector catches "AI pretending to be a person," and we also check the file for provenance (C2PA-style AI markers). Still indicative only — heavily edited real photos can cause false positives.