Google has launched a public website that allows users to check images, video, and audio for invisible AI watermarks. The tool identifies markers applied by Google and its technology partners, including OpenAI, ElevenLabs, Kakao, and Nvidia.
The SynthID technology embeds signals directly into content, such as pixels for images and video or waveforms for audio. According to Google, Gemini models have applied these watermarks to over 100 billion images and videos, as well as 60,000 years of audio content.
Google DeepMind introduced SynthID in 2023, initially focusing on images before expanding to text, video, and audio. A limited version of the detector was released in May 2025, while integration into Google Search and Chrome followed in May 2026.
The company is also incorporating the C2PA industry standard to provide additional provenance data for files. Future developments include the integration of SynthID into Apple's iOS 27 and the launch of SynthID Bio for protein sequences on September 30, 2026.
