Deepware Scanner deepfake detector

The archived scanner page offers one action: "Place a video link or upload a video", with a SCAN button, a BETA label and a note that "By submitting data, you are agreeing to Terms of Services and Privacy Policy" (capture of 8 August 2026, read 9 November 2026). The FAQ sets the limits: "a maximum limit of 10 minutes for each video", links from "YouTube, Facebook, and Twitter", and the resolution recommendation above (deepware.ai/us/faq, read 9 November 2026). No account, quota or monthly cap is published for the web scanner, and no price is published for anything. The FAQ describes three access routes: "We provide Deepware Scanner in web-platform, API, and SDK forms. You can either use our website, our API key, or in an offline environment via SDK. Contact us to learn more."

Retention is governed by terms that treat every upload as a shared sample. The terms of service read: "you give Deepware.ai (and those we work with) a worldwide, royalty-free, irrevocable and transferable license to use, edit, host, store, reproduce, modify, create derivative works, communicate, publish, publicly perform, publicly display and distribute all content contained in the Sample", and "if you submitting any Sample, the Sample is immediately shared for review by the Service's Partners, and the resulting intelligence report is shared with you and, and with the Partners, which use the results to improve their own systems" (deepware.ai/us/terms-of-services, undated, read 9 November 2026). The same page tells users to "only upload suspicious samples that you wish to publicly share" and, for anyone who does not, "do not send it/contribute it to the service". The privacy policy states no period: "We will retain your Personal Data only for as long as is necessary for the purposes set out in this Privacy Policy" (deepware.ai/us/privacy-policy, undated, read 9 November 2026). The retention rank field is therefore empty and the clauses are quoted on the card.

Ownership is stated in two footers. deepware.ai gives "Zemana Doo, Splitska 6 / 71000 Sarajevo, Bosnia and Hercegovina" with a 2022 copyright, and the privacy policy says "Zemana operates https://deepware.ai" (deepware.ai/us, read 9 November 2026); the archived scanner footer gives "Bilkent Cyberpark Cankaya, Ankara, Turkiye" with a 2021 copyright. The about page says "we have open-sourced our current deepware scanner", but the GitHub organisation the site links to (github.com/deepware) held only a face-detection library, dface under the BSD-2-Clause licence, and a Pix2Pix implementation on 9 November 2026, both last pushed in February 2021, so the card does not list the tool on the open-source view.

Deepware publishes no accuracy figure. Asked "What is your detection accuracy?", the FAQ answers "Please visit our GitHub page to learn more about our detection accuracy with detailed information" (deepware.ai/us/faq, read 9 November 2026), and the linked GitHub organisation holds no scanner repository and no figures. The nearest statement to a claim is a caveat: "Can I trust your result 100%? Deepfakes are not a solved problem yet. So our results are stating the chances of a specific video can be whether deepfake or not." No false positive rate is published, and no statement is made about compressed, re-encoded or edited video, so the humanized-text field, which does not apply to video, is empty.

The vendor's own description of the threat is on the home page: "Cybersecurity faces an emerging threat generally known as deepfakes. Malicious uses of AI-generated synthetic media, the most powerful cyber-weapon in history is just around the corner" (deepware.ai/us, read 9 November 2026). The deepware.ai footer carries a 2022 copyright, and the scanner is still labelled BETA in the August 2026 capture.