Detecto
Image

AI image detection, on the file you uploaded.

Detecto returns a generation likelihood, a region-level heatmap, surface-artifact flags, and a metadata-integrity check. Designed for newsroom verification desks, citizen-contribution pipelines, and PR teams clearing supplied imagery.

Try a scanHow calibration works
Try a scan

Drop a file here, or pick one

This is a quick demo, files are not saved. Sign in to keep your reports.

1 credit per image

No account needed. Anonymous scans are rate-limited per IP.

Built for the people who review other people’s content

  • Editors at independent magazines
  • University writing centers
  • Newsroom verification desks
  • PR review teams
  • Content QA leads at agencies
Accepted inputs

Files you can drop in.

JPG, PNG, WebP. Up to 10 MB per file. URL ingestion is out of scope for v1.

Citizen photos
Submitted photos from contributors, witnesses, and the public.
Press-supplied imagery
Press release imagery, supplied portraits, vendor-provided product shots.
Editorial composites
Stock-like images used in editorial composites and explainer pieces.
Reviewer-flagged uploads
Images escalated from review for an additional pass before publication.
What we measure

Signals exposed in every image report.

The same vendor-agnostic feature labels every reviewer sees, regardless of which generator produced the input.

Aggregate
Generation likelihood
Aggregated likelihood across the surface-artifact ensemble. Reported as a percentage.
Region-level
Region heatmap
Per-region likelihood grid surfaced as a heatmap reviewers can read at a glance.
Surface
Surface artifacts
Hands, skin, lighting consistency, micro-texture coherence flags.
Metadata
Metadata integrity
EXIF, file structure, and re-encode trace, flagged when the file claims a provenance the bytes don’t support.
Frequency
Frequency residual
High-frequency residual patterns characteristic of common diffusion-decoder pipelines.
Calibration
Confidence variance
Variance across resampled detection passes on the same input.

Honest limitations on image detection

A detection score is probabilistic, not proof. Treat the report as evidence to discuss, not a verdict to publish.

  • Heavy re-encoding (compression, screenshot of a screenshot) erodes metadata signals.
  • Newer generation models reach the surface-artifact ensemble before the calibration plate updates.
  • Heavily edited human photos can produce artifact patterns that look generation-typical.
  • Default verification posture: detection is one input among several, not the verification workflow itself.
  • No URL ingestion in v1 │ file-based only.
For verification desks

Run the file through Detecto before the article lands.

Your reports save privately to your dashboard, shared only when you choose.

Get startedFor publishers