FakeRadar

FakeRadar is an online tool designed to detect AI-generated images and deepfakes by analyzing uploaded images or short videos. It addresses the need to identify manipulated or synthetic content by providing users with a rapid assessment of whether a file has been generated or altered by artificial intelligence. The service delivers a clear report indicating the presence of AI signals, manipulation scores, confidence levels, and likely model families involved in content creation.

Users can upload files in JPG, PNG, WEBP, GIF, or short video formats, with a maximum size of 10 MB. Once uploaded, the original file is deleted from the servers after analysis, ensuring privacy. FakeRadar employs advanced AI models that scan content across multiple layers, evaluating metadata inconsistencies, pixel artifacts, model signatures, and other signals such as light inconsistency. The analysis includes multi-engine techniques, referencing methods like ELA, FFT, and C2PA, and provides results in under 10 seconds on average, with an advertised accuracy of 94%.

The generated report summarizes the findings, including scores for AI signal strength, manipulation, overall confidence, and details about possible model families, such as the Midjourney family. Reports can be saved or shared via a single link, making it convenient for users to distribute results. FakeRadar emphasizes that its results are analytical signals rather than definitive proof and are not intended to serve as legal documents.

The platform is web-based and offers a free tier for the first three analyses. No information about further pricing, licensing, or targeted user groups is provided beyond the general context of image and video authenticity verification.