AI-Generated Image Signal
Generative-model fingerprints summarized as a 0–1 strength with a risk band, so you see how strong the evidence is — not just a score
EVIDENCE-AWARE AI CHECKER
Is this AI generated image?
Upload an image and get an evidence-aware answer in seconds — AI-generated and face-manipulation signals, free, deleted right after the check.
- Image deleted right after the check
- Uploads are not used for training
Your images stay private
Not stored permanently on our servers
Not shared or used for training
Your content remains yours
Results can be uncertain
Always review and use with context
Core signals
AI-Generated and Face-Manipulation Signals
Every report centers on two independent signals: is the image generated, and was a face manipulated
Face-Manipulation Signal (Deepfake)
An independent signal for manipulated faces. It only appears when a face is present and never changes the main risk band on its own
What It Cannot Do
Stylized, illustration-like generations can score lower and land in the uncertain band. We show that honestly instead of forcing a guess
IMAGE REPORT · Demo data
Inside an Isthisagenerated Report
— Evidence, Not Just a ScoreIt is more than a score: a risk band, the independent signals behind it and the limits that shape the conclusion
This is a structural preview. The band and values are illustrative and do not represent a real image's detection result.
AI-Generated Risk Band
Uncertain
Needs a second look
The band summarizes signal strength under current input conditions. It supports judgment and is not a final real-or-fake conclusion
Report Coverage
Complete
The report states which signals were used and which input conditions may affect the result
Input Conditions
- Source: uploaded image
- 1024 × 1024 · JPEG
- 1.2 MB · file readable
AI-Generation Signal
Strong enough to flag, not enough to conclude — stylized images often land here
Face-Manipulation Signal
No clear face in this image, so the deepfake signal is not part of the conclusion
What the Report Explains
- AI-generation signalIncluded
- Generator family cluesIncluded
- Face-manipulation (deepfake) signalNot included
- Source metadata (EXIF)Not included
Result Limits
- Stylized or illustration-like images can score lower than photorealistic ones
- Compression, re-encoding and screenshots can weaken available signals
- A signal not included in the conclusion is not treated as a normal result
Recommended Verification Steps
- 1Do a reverse image search to find the earliest posting
- 2Check the posting account's history and credibility
- 3Ask the sender for the original file when it matters
What You Receive
- A risk band, not a bare score
- Independent AI and face signals
- Limits and next-step checks
Demo data only illustrates the report structure and does not represent real detection capability
How it works
How to Check If Something Is AI Generated
in Three StepsFrom upload to report, Isthisagenerated turns complex detection output into information you can understand and act on
- step_01
Drop in an image
Drag, paste or pick a JPEG, PNG, WebP or AVIF up to 10 MB
- step_02
Scan for AI signals
The checker looks for generative-model fingerprints and face-manipulation traces
- step_03
Read the evidence
Get a risk band with the signals behind it — not a bare score
Use cases
What People Check With Isthisagenerated
Before you share, buy or publish — look at the signals and the limits first
Images from chat & social
A friend forwards a wild “real photo”. Check the AI-generation signal before you pass it on
Open the image detectorMarketplace & profile photos
Too-perfect product shots and too-clean portraits. Get a read before you trust them
Open the image detectorNews & viral images
Before you share that shocking photo, see the signals and the limits behind them
Open the image detectorPrivacy by default
Your Image Is Deleted Right After the Check
We make the data lifecycle visible: your image goes away right after the check, and the full privacy and deletion policy is always available
Data lifecycle
- Your uploaded imageDeleted right after the check
- Structured text reportKept for 30 days by default
- Training useNever without explicit authorization
A promise you can check
The report shows available signals and limits. Image deletion, report deletion and training authorization remain separate controls.
Read the full privacy & deletion policyDecision Support, Not a Conclusion
Isthisagenerated results are decision support only. They are not a factual finding, legal evidence or the sole basis for action. Read the signals and limits together with provenance and human review
Help & boundaries
Is This AI Generated? Common Questions
A few clear answers about what Isthisagenerated can show, how to read it and how your image is handled
What does Isthisagenerated check?
Evidence-aware checks for images: an AI-generation signal, an independent face-manipulation (deepfake) signal, a risk band and the limits behind it. It tells you what the signals show — it does not issue a final real-or-fake conclusion.
Can a checker confirm an image is AI-generated beyond doubt?
No. A high signal means generative fingerprints are strong, which is still not a final conclusion. Use the report as decision support, verify the source and keep humans in the loop for anything important.
How accurate is it?
Genuine photos rarely trigger false alarms in our regression set. Stylized generations (illustration-like output) can score lower and fall into the “uncertain” band — we would rather say uncertain than guess.
Which AI generators does it catch?
Signals are strongest for the GPT-Image/ChatGPT, Midjourney and Stable Diffusion families, and they improve as generators evolve.
Is my image stored or published?
No. Your image is deleted right after the check, the report is visible only to you, and report pages are marked noindex so they do not surface in search.
Is it free?
Yes during the beta — no card needed. Image checks do not consume any video minutes.