Stable Diffusion Detector
Stable Diffusion has a peculiarity worth knowing: run locally through AUTOMATIC1111 or ComfyUI, it frequently writes the entire generation recipe into the file — prompt, seed, sampler, model. People share these images without realising the instructions are still inside.
What a local Stable Diffusion install leaves behind
Hosted Stability API output carries C2PA credentials. Local installs do not — but they habitually store something far more revealing in plain text.
The whole recipe, in the file
AUTOMATIC1111 and ComfyUI commonly embed the positive and negative prompt, seed, sampler, step count, CFG scale and model hash. Not a subtle trace — a full confession, readable in our free EXIF viewer.
ComfyUI workflow blobs
ComfyUI can store the entire node graph as JSON inside the PNG. If it is there, there is no ambiguity left to resolve.
No C2PA on local runs
Stability's hosted API embeds Content Credentials; a local install does not. So metadata may be either extremely revealing or completely absent.
Model-specific rendering
Thousands of community checkpoints and LoRAs produce distinctive looks, but all of them share the underlying diffusion statistics our models are trained on.
Anatomy under load
Hands, crowds and complex poses remain the classic failure points, especially on older checkpoints still in wide use.
Upscaler traces
SD workflows almost always end with an upscaler, which leaves its own frequency signature — detectable even after the metadata is stripped.
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.
If the file carries no generation parameters, Start with our AI image detection tool — 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.
Who checks for this
- Art platforms — Enforcing AI-disclosure rules.
- Clients — Verifying that commissioned work is original.
- Researchers — Recovering the prompt and parameters behind an image.
- Moderators — Community rules on generated submissions.
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.
Start with the metadata on this one. If a prompt and seed are sitting in the file, you have your answer in seconds and do not need us at all — we would rather tell you that than sell you a check you don't need.
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.