Dating & marketplace safety
Verify a profile photo
Run the picture through the AI image detector before you trust it.
- Shareable report link for moderators
Multi-signal forensic screening
AI Image Detector
Is this picture AI-generated? Get evidence, not a guess.
Upload an image and screen it with four independent AI-detection models cross-checked by a published weighted-vote rule. Graded conclusions with the evidence laid out — never a made-up percentage.
- Four detection models, published weights — every vote shown by name
- Independent engines cross-checked — disagreement is reported, not hidden
- Batch AI image detector — up to 50 images at once, triage summary + CSV
- 4 welcome credits on sign-up — your first full report is free
Intake · Image screening
4 credits per report · first one free
See a real report
Evidence laid out, not a lone percentage.
Real reports from this AI image detector — the exhibit, each engine’s vote and the visual inspection, side by side. Click any one to read it in full.
The signals
One AI image detector,
multiple independent witnesses.
Four independent detection models under a published weighted-vote rule — plus a trump card: deterministic file evidence that, when present, decides outright.
Credentials & metadata trump card
C2PA credentials and generator self-tags are checked first. They only speak when certain — and then they decide, over every model.
Sightengine GenAI
Commercial AI image detector engine with per-generator classification — Midjourney, Flux, Imagen and more.
Winston AI
An independent second engine — cross-engine agreement is what makes the consensus mean something.
Was It AI
A third independent commercial engine, credited by name in every report where its vote appears.
Open-source cross-check
SAFE (KDD 2025) runs on our own hardware as an independent weighted vote — fully auditable.
Every signal. One verdict.
Analyze an imageRun an AI-detection engine? We're always looking to add independent engines to the report — every result is attributed by name, never blended into a single score. Write to hello@aiimagedetector.info.
Coverage
The AI image detector for every major generator.
This AI image detector classifies against the fingerprints of current image models across four independent engines.
Midjourney
v5–v7 styles
DALL·E
OpenAI
GPT-image
ChatGPT
Stable Diffusion
and fine-tunes
Flux
and fine-tunes
Google Imagen
Gemini
Adobe Firefly
Adobe
Ideogram
text-heavy
GAN faces
synthetic portraits
New generators ship monthly — exactly why four independent engines vote inside this AI image detector instead of one model guessing alone.
How it works
From upload to AI image detector verdict in seconds.
Every signal runs in parallel.
- Step 01
Upload the untouched original
Drop, browse or paste — the AI image detector analyzes your exact bytes and records the SHA-256.
- Step 02
Every model runs in parallel
Four detection models vote with published weights. A dead engine never blocks the report — its credit refunds automatically.
- Step 03
A graded conclusion you can defend
Five grades from AI-generated to real, disagreement reported honestly and named.
What you can check
Any image where real or fake matters.
Feed the AI image detector a JPEG, PNG or WebP — or fifty in a batch.
Profile & dating photos
Screen a face with the AI image detector first.
Digital art & illustrations
Keep human-made galleries human.
Product & listing photos
Catch generated product shots before you pay.
News & viral images
Through the AI image detector before you share.
Social media re-posts
Stripped files work too — pixels still testify.
Screenshots & memes
Weaker evidence, honest hedging.
Who uses it
The AI image detector for pictures that have to be true.
This AI image detector tells you when to trust, when to verify, when a human should look.
Educators
Evidence, not accusations
A graded AI image detector report can carry an integrity conversation; a bare percentage cannot.
- "Recommend review" language, never verdicts
Artist communities
Keep human spaces human
A defensible AI image detector instead of gut feel and comment forensics.
- Batch triage built for mod queues
Editorial & research
Provenance before publishing
A documented AI image detector pass — four models, one weighted consensus.
- SHA-256 for chain-of-custody notes
Why this one
A typical AI image detector, and this one.
A typical detector
AI Image Detector
One model, one percentage
Every signal weighed into one graded verdict — sources named
Hides engine disagreement behind one number
Reports conflict out loud — “can’t say for sure” is an answer
Charges the same when an engine fails
A dead engine refunds its credit automatically
Free tier is a teaser
The trial runs the full AI image detector — every signal
Pricing
Every feature is free. Pay for volume.
Sign up and your first AI image detector report is free — plans are for volume.
Free
$0on sign-up
4 credits — 1 full report
The whole product, no card required.
- The full AI image detector, every signal
- Shareable report links + forensic PDF download
- First full report on us
Most popular
Pro
$19/ month
300 credits / month
For people who screen images all day.
- Everything in Free, at volume
- 4 credits per report — 75 reports monthly
- Batch screening — 50 images, CSV export
Team
$59/ month
1500 credits / month
For trust & safety and moderation teams.
- Everything in Pro
- Shared workspace history
- Priority support
A report costs 4 credits today. That number tracks the engine pool: when we add detection engines to the report, the per-report cost may go up. Changes are announced in advance and never apply to reports you have already run.
FAQ
Straight answers about this AI image detector.
What this AI image detector can and cannot tell you.
Ready when you areRun the AI image detector
Run the AI image detector
in the next 30 seconds.
The full AI image detector — first report free on sign-up.