DeepMark

DeepMark is a verification service for AI generated media and live AI voice. It embeds an imperceptible, machine readable identity into audio, image and video as the content is produced, so the output looks and sounds the same to a viewer or listener. The signal can be detected later and resolved to a clear pass or fail result, with signed evidence behind the outcome. The platform serves two purposes. Content marking covers audio, image and video that must carry a provenance signal, and the vendor says it is built to support transparency obligations under Article 50 of the EU AI Act. Voice agent authentication lets a receiving party confirm that an incoming call came from an authorized AI agent, and the vendor says the signal survives VoIP, codecs and the telephone network. The vendor names banking, contact centers, healthcare, insurance and production companies as target sectors. Access starts with a request for a demo. The vendor's site does not publish plan prices, and it does not describe a trial or self service tier. The company says its audio watermarking work began in a university research lab in Montenegro and that it is backed by Y Combinator's Fall 2026 batch. It states that its audio watermarking technology is commercially licensed and in production use.