Google has opened its SynthID Detector to everyone, making the tool available globally in English. Announced by Pushmeet Kohli, VP of Science and Strategic Initiatives at Google DeepMind, the expansion moves the detector beyond the journalists, media professionals, and researchers who had early access.
The web tool, hosted at synthid.com, lets users upload an image, video, or audio file to check whether it carries a SynthID watermark — an imperceptible signal embedded in media generated or altered by participating AI tools.
What it covers — and what it does not
The detector recognizes watermarks from Google and its partners, including OpenAI, NVIDIA, and Kakao, with Apple support listed as coming soon. Google's own Gemini, Imagen, Lyria, and Veo models apply the mark, as do tools such as Flow, ProducerAI, and Vids.
It is not, Google stresses, a general AI detector. The site states plainly: "This is not a general AI detector. We can only detect media from the above companies, who have adopted SynthID technology." Other companies, such as Meta and Microsoft, use their own watermarking standards, and some open-weight models carry no watermarks or metadata at all.
A signal, not a verdict
A positive result indicates that the media was made or edited with Google AI — or, for partner content, the AI of a participating company — wording that does not distinguish a file generated from scratch by AI from one that was merely touched up with AI tools. Google's sample result also notes that the file may have been edited further since the mark was applied.
A negative result is even less conclusive. The absence of a SynthID watermark means only that the detector did not find a mark it recognizes — not that the file is authentic, untouched, or made by a person. Watermarks can be stripped or degraded in some cases, and media from generators that have not adopted SynthID will simply not register.
Access limits
Using the detector requires signing in with a Google, OpenAI, or Apple account. Google says each user gets "a daily usage quota of approximately 10 image, video, and audio checks." Google engineers have previously said that limiting access helps prevent people from using repeated checks to develop watermark-removal tools.
The scale behind the system is large: Google says it has watermarked more than 180 billion images and videos, along with 240,000 years of audio content, since introducing SynthID in 2023. Built-in verification in Search, the Gemini app, and Chrome now handles more than one million requests a day.
What is confirmed, what is unknown
Confirmed: the detector is now publicly available globally in English; it checks images, video, and audio for SynthID marks from Google, OpenAI, NVIDIA, and Kakao; and both positive and negative results are limited signals rather than proof of how a file was made.
Unknown: how the public detector performs under independent testing — the available material includes no audited false-positive or false-negative rates — and how reliably watermarks survive compression, cropping, or deliberate removal attempts.
Still awaited: a date for Apple's announced support, and any indication of languages beyond English.
Reporting: Google, Search Engine Journal, TechCrunch, The Next Web, Ars Technica, Eurovision News Spotlight.