— Image forensics · calibrated verdicts
Forensic-grade AI image detection.
A detector fine-tuned for current generators scores your image and returns a verdict at a capped false-positive rate — 5% for triage, 1% for evidence — with thresholds calibrated per condition and every limit stated on the report, not buried in a footnote.
Scored in memory · never stored · verified chain of custody
Forensic report
VF-9E41C7A2
thr 0.725 @ ≤1% fpr · detection rate 43% under social chain
primary + supporting signals · robustness panel · limits
8,064
benchmark images
8
real-world conditions
386,072
scored evaluations
6
detectors on the calibration grid
— How it works
Three steps between an image and a defensible answer
01 · Hash
Fingerprinted in your browser
Your browser computes the SHA-256 of the exact bytes before upload. The returned report must hash the identical bytes or the client rejects it — chain of custody is enforced, not just printed.
02 · Score
A calibrated primary, corroborated
A peer-reviewed detector (CVPR 2025 lineage), fine-tuned on 2024–26 generators and pinned by checkpoint hash, drives the verdict; independent supporting signals corroborate it.
03 · Decide
A verdict with a false-positive budget
Thresholds are calibrated so at most 5% — or 1% — of real photos would be flagged. The report states detection power under your image's condition.
— Why Vestige
Confidence, including its limits
Detection tools love to quote lab-bench numbers. We publish what survives contact with reality — recompression, messenger re-encodes, screenshots of screenshots.
- The same detector that scores 0.93 AUC catches ~55% of AI images under a social-media re-encode chain when false positives are capped at 5% — and ~43% at the strict 1% standard.
- No tool on the market clears the ~90% evidentiary gate under realistic conditions. Anyone implying otherwise is selling you a liability.
- So every report quantifies exactly what a flag — or its absence — is worth for your image, under its estimated condition. That number is your defensible ground.
— Who it's for
Built for the lab. Useful far beyond it.
For specialists
Digital forensics & legal
FPR-capped verdicts, chain of custody, and the caveats a defense expert will ask about — printed on the report before they ask.
Learn more →
For enthusiasts
OSINT & fact-checking
Triage viral images with calibrated scores instead of vibes. Built for the recompressed screenshots you actually encounter.
Learn more →
For private users
Personal checks
A dating photo, a marketplace listing, a too-perfect “family emergency” picture. Get a straight, plain-language answer.
Learn more →
— In development
Next on the same engine
For law firms, PIs & HR investigations
Evidence review for legal teams
Court-ready reports on disputed images, with error rates a cross-examination can’t shake.
For insurers & TPAs
Claim-photo verification
Generation, splice, and recycled-photo signals on claim photos — after the re-encode chain.
For journals & research-integrity offices
Scientific-image integrity
Batch figure screening for manipulation, duplication, and generation — humans decide.
— The report
Every claim on it is auditable
SHA-256 of the exact bytes analyzed. Thresholds tied to a named calibration run. A hash of the report itself. Built to be attached to a case file, not a slide deck.
Privacy, in three commitments
- Your image is scored in memory and discarded — never written to disk, never logged, never kept.
- Chain of custody is enforced: your browser hashes the bytes before upload and rejects any report that hashed different bytes.
- There is nothing to delete server-side. The report is returned to you and exists only where you keep it.
— FAQ
Fair questions, straight answers
Can you tell me definitively whether an image is AI?+
No — and be suspicious of anyone who says yes. We return a calibrated statistical verdict at a false-positive rate you choose, with the detection rate and every known limit printed on the report. That is the strongest honest claim current science supports.
What kinds of images work?+
JPEG, PNG, and WebP. Re-shared, recompressed, screenshotted images are exactly what the pipeline is calibrated for — our benchmark includes a “screenshot of a WhatsApp image” condition because that is what real cases look like.
Do you keep my images?+
No. Your image is scored in memory and discarded — it is never written to disk, never logged, and never used for training. The report is returned to you and exists only where you keep it.
What does the free analysis include?+
The complete forensic report — calibrated verdict, primary and supporting detector scores, robustness panel, limits, and chain-of-custody hashes. One image per account, then pay-per-image credits or an API plan.
Analyze your first image free
Create an account, upload one image, and read the full forensic report — thresholds, limits, hashes and all. No card required.