Adobe Firefly Detector
Adobe built Content Credentials into Firefly from the start — every image where all the pixels are generated gets a signed manifest recording the model and the actions taken. Adobe is the strictest of the big generators about this, which makes Firefly output unusually traceable.
What Adobe writes into the file
Firefly is the clearest case in this category — as long as nobody has laundered the file through a screenshot.
Automatic Content Credentials
Adobe applies C2PA credentials to every image where 100% of the pixels come from Firefly, recording which model was used and the processing actions taken.
Generative Fill is a different animal
When Firefly edits part of a real photo, only that region is synthetic. The file may be marked as edited rather than generated — and the region itself shows up in our manipulation analysis.
Credentials survive inside Adobe
Moving between Adobe apps preserves the manifest. Exporting elsewhere, screenshotting or re-saving usually destroys it.
Software field
Even without a full manifest, Adobe applications commonly leave their name in the metadata — visible in our free EXIF viewer.
Stock and commercial context
Firefly is trained for commercial safety, so its output turns up in marketing and stock contexts where disclosure obligations often apply.
Generation statistics
Underneath the marketing, it is still a generative model: flat noise, even detail distribution, no capture parameters.
How the analysis works
Upload the original file
Full resolution, straight from the source. Screenshots and re-saved copies destroy most of the evidence we measure.
12 independent methods run in parallel
Three neural networks look at the image as a whole. Nine forensic methods examine it as a signal: frequency (FFT) analysis, noise pattern, pixel-level statistics, EXIF metadata, Error Level Analysis, texture coherence, colour distribution, Benford's Law and edge detection.
You see the breakdown, not just a number
Every method reports separately, so you can tell whether the verdict rests on one weak signal or on nine that agree.
When Content Credentials have been removed, Start with our AI image detector — it covers any image type and runs the same twelve methods.
How much each method counts
Most detectors give you one number from one model. We run twelve independent analyzers and weight them — three neural networks carry the majority, but nine forensic methods examine the file as a signal and can overrule them. Here is the exact weighting we use in production:
| Method | Weight in final score |
|---|---|
| Neural network #2 | 25% |
| Neural network #1 | 15% |
| Neural network #3 | 15% |
| EXIF metadata | 10% |
| Frequency (FFT) | 6% |
| Benford's Law | 6% |
| Noise pattern | 5% |
| Error Level Analysis | 4% |
| Texture coherence | 4% |
| Edge detection | 4% |
| Pixel statistics | 3% |
| Colour distribution | 3% |
What people actually upload
These are our own numbers, not an industry estimate — every check run through FakeSec since launch. The split is closer than most people expect: roughly four in ten uploads really do turn out to be generated.
Where it comes up
- Agencies and clients — Deliverables that were specified as photography.
- Stock libraries — Submissions with AI-disclosure requirements.
- Compliance — Advertising rules on synthetic imagery.
- Publishers — Editorial policies on generated illustration.
Accuracy — and where it breaks down
On our internal benchmark of 240,000 images (120,000 AI-generated, 120,000 real) the combined score reaches up to 98% on unedited, full-resolution files. That number drops, sometimes sharply, in four situations — and you should know them before trusting any result:
- Heavy compression. Social platforms re-encode everything. A photo pulled from Instagram has already lost much of its noise signature.
- Screenshots. A screenshot is a new image of an image. Original metadata and compression history are gone.
- Stylised content. Flat colour and clean lines remove the texture and noise evidence forensic methods depend on.
- Post-processing. Added grain, filters, upscaling and manual retouching all mask generation traces.
No detector on the market is 100% accurate — ours included. Treat a high score as strong evidence, not as proof. When a decision matters, look at which of the 12 methods agreed.
For Firefly specifically, metadata is the first and best test — Adobe is unusually consistent about applying it. If the credentials are missing, that most often means the file was exported or screenshotted, not that it is authentic.
Frequently asked questions
Related tools
Check your first image now
Create a free account and get 3 checks. Full breakdown across all 12 methods, no card required.