Google opens the SynthID Detector to everyone: check any image, video or audio for hidden AI watermarks
Google's SynthID Detector is now open to everyone, globally: upload any image, video or audio file and check whether it carries an AI watermark from Google, OpenAI, NVIDIA or Kakao. Apple support is coming.
Google has opened its SynthID Detector to everyone, ending more than a year of gated access for media professionals and researchers. In a post on The Keyword, written by Pushmeet Kohli, VP of Science and Strategic Initiatives at Google DeepMind, the company says the portal at synthid.com is now available globally in English starting today — October 7, 2026.
Anyone can upload an image, video or audio file and learn whether it carries the imperceptible watermark Google calls SynthID — not just from Google’s own models, but from partner AI systems too: OpenAI, NVIDIA and Kakao are covered, with Apple support listed as coming soon. “As AI tools become better and easier to use, having clear context about the content you encounter online helps you make informed decisions about what to trust,” Google writes.
The scale is already enormous. Since launching SynthID in 2023, Google says it has watermarked over 180 billion images and videos, along with 240,000 years of audio content. The Detector joins built-in verification features in Search, the Gemini app and Chrome, which the company says now handle more than 1 million requests daily.
What it checks — and what it doesn’t
The portal is explicit about its limits. Its FAQ states: “This is not a general AI detector.” It can identify only media from companies that have adopted SynthID technology, and it says work continues on extending partnerships to cover more generated media. A negative result does not prove a file is human-made — the content may have come from a model without SynthID, or heavy modification may have degraded the signal. The FAQ also notes rare cases of false positives, and recommends uploading the highest-quality file available.
The watermark itself is designed to survive the real world: for images and video it is embedded at the moment of creation, in the pixels rather than metadata, and is meant to remain detectable after cropping, filters, frame-rate changes and lossy compression. Supported formats include JPG, PNG, WEBP and HEIC images; MP4, MOV and WEBM video; and WAV, MP3, OGG, FLAC, AAC and M4A audio.
The industry is lining up behind watermarks
The Detector’s first public version appeared at Google I/O in May 2025 as a verification portal for content made with Google’s own AI tools, behind a waitlist for journalists and researchers. By May 2026, Google said OpenAI, Kakao and ElevenLabs were extending SynthID watermarking across more of their AI-generated content, and NVIDIA had partnered to watermark video from its Cosmos models. Google also launched an AI Content Detection API on its Gemini Enterprise Agent Platform for business users.
That momentum is the point: a watermark detector is only as good as the share of AI content that carries a watermark. Getting OpenAI — the biggest generator of consumer-facing AI media on earth — and Apple on board turns SynthID from a Google feature into something closer to a default industry check. The company says Apple support is next.
Why it matters
For newsrooms, advertisers, courts and anyone trying to sort real from synthetic, a one-click public checker from Google is a genuine step forward — provenance tools only work if ordinary people can reach them. But the boundary conditions matter as much as the launch: a clean result is not proof of authenticity, a watermarked file only proves the file came from a participating model, and anything from a lab that hasn’t signed on is invisible to the tool. SynthID is one component of a wider provenance strategy — worth using, and worth understanding the limits of.
Sources: Google (official announcement, Oct. 7, 2026); Unite.AI; WebProNews.