Google has opened its SynthID verification website to the public, allowing anyone to check images, video and audio for the company’s invisible AI watermark. The tool was previously limited to journalists, researchers, and media professionals.
The SynthID website does not identify every piece of AI-generated media. It looks for SynthID watermarks embedded by supported models and products, so a positive result indicates compatible AI tools were involved in creating or modifying the file.
SynthID supports images, video and audio
The detector accepts common image formats, including JPG, PNG, WEBP, AVIF, HEIC, TIFF, and GIF, along with MP4, MOV, and WEBM video files. Supported audio formats include WAV, MP3, OGG, FLAC, AAC and M4A.
Google introduced SynthID in 2023 and now uses it across products and models including Nano Banana, Veo, Lyria, Gemini, Flow, ProducerAI and Vids. OpenAI also supports SynthID and operates its own verification service, while Nvidia and Kakao are among the other companies using the technology.
Verification is expanding across Google products
Google has built SynthID checking into Gemini and Chrome and says users make about 1 million verification requests each day.
Watermark detection still has limits, particularly after media has been edited. A Reuters analysis of Meta’s detection technology, for example, found that cropping could prevent some of the company’s own AI-generated images from being recognised.
Google’s public SynthID site therefore provides a useful verification signal, but not a universal test for whether content was created with AI.
