Terms of use
Effective October 2, 2026
What ImageCan is
ImageCan analyzes an uploaded image with a fixed set of signal modules (EXIF, perceptual hash, C2PA, an AI-generation classifier, and vision-model triage where available) and returns a per-signal breakdown plus a composite verdict: camera-like lean, generated-or-edited lean, or insufficient signal. For an uploaded video, the image classifier reads twelve frames spread across the clip; that reading can support a generated-or-edited lean and can never support a camera-like lean, so a video that is not flagged is reported as insufficient signal, not as authentic. Each signal states its own confidence; several are explicitly informational and are excluded from the composite.
No professional advice or warranty
The service is experimental and provided as available. Its heuristic signals are known to be defeated by re-compression, resizing, and adversarial editing, and are known to misjudge both real and generated images. Honeycutt AI Labs LLC does not warrant uninterrupted operation, accuracy, or fitness for a particular purpose, and does not warrant that a result correctly identifies whether an image is authentic, edited, or generated.
Acceptable use
Upload only images and videos you are entitled to share for analysis. Do not use ImageCan's output as the sole basis for an accusation against a named person, or as a substitute for a forensic or legal determination. Image uploads over 16 MiB or 20 megapixels, video uploads over 32 MiB, and files in an unsupported format are rejected.
Capacity
ImageCan runs on a single shared host with a concurrency limit of one analysis at a time. A busy service returns an error asking you to retry; it does not queue your request in this release.
Changes
Material changes will be posted here with a new effective date.