How an image check works.
The three tools use the same IsItAI classification API; their different contexts do not imply different forensic capabilities.
1. Validate and prepare the image
We check file signatures, container structure, dimensions and size before dispatch. JPEG APP1/comments, PNG text/EXIF/time chunks and WebP EXIF/XMP are removed; GIFs use their first frame. This is not a full image decoder or a guarantee that all identifying content is removed. Visible faces, text and contextual details can remain. HEIC and animated WebP are unsupported.
2. Make one provider request
The server reserves a credit and submits the prepared image to IsItAI. The API credential stays on the server. Retries do not automatically repeat an inference whose outcome is unknown.
3. Interpret the output
IsItAI returns AI or Human and a confidence value. We display the selected class and its score without turning Human into a claim that an image is authentic. A suggested generator, when returned, is a model suggestion rather than verified provenance. The provider does not expose a pinned model revision in this response.
4. Understand the limits
We have verified the API transport using project-created fixtures. We have not independently established accuracy, calibration or reliability across current generators, art styles and edited photos. There is no manipulation heatmap, verified C2PA trust chain, face identity check or Photoshop detection guarantee.
5. Keep or remove your result
Your image and report are private by default and expire after 24 hours. Export a report if you want to retain it. Sharing is explicit, score-only, revocable and unlisted. The provider’s own data handling is separate.
Provider details: IsItAI privacy policy · IsItAI terms.
Imagefake is independently operated under the name Imagefake Team. Contact support@imagefake.com for support, billing or privacy questions.
Contact Imagefake