Upload an image (JPG, PNG, WEBP max 4MB) to analyze synthetic visual signals and extract C2PA credentials.
Free limits and planned Pro access
Drag & drop an image here or click to browse
JPG, PNG, WEBP — Max 4 MBAI analysis: image sent to Hive. C2PA provenance: checked locally in your browser. A score is not proof.
The classifier looks for pixel-level artifacts left by diffusion models and GANs, and estimates the closest known generator (Midjourney, DALL·E, Flux, Stable Diffusion, Firefly and others). In parallel, your browser reads any embedded C2PA Content Credentials.
Upload the original file whenever possible. Screenshots, heavy compression, resizing and filters remove the signals detectors rely on.
Yes. Re-compressing or screenshotting images modifies pixel distribution and can lower signal accuracy.
JPG, PNG, and WEBP files up to 4MB are supported directly.
Only partially. The model is designed for fully generated images; local AI edits (inpainting) on a real photo may not be flagged.
Batch analysis, higher limits, PDF reports and an API for platforms. Get notified at launch.
Use an address you can access. This form saves your interest; it does not verify your mailbox. Deletion on request.