What opened

A year ago at Google I/O, the Detector was handed to a small group of journalists, media professionals, and researchers. This week it went public: a browser upload at synthid.com, live globally but English-only for now. Sign-in is required. Google is also pushing verification into places you already are. Checks are built into Search, the Gemini app, and Chrome, where the company says more than a million verifications run every day.

Who's covered, and who isn't

Detection only works for media that carries SynthID. That means Google's own models, Gemini, Veo, Lyria, Nano Banana, and the tools around them: Gemini, Flow, ProducerAI, Vids. Partners OpenAI, NVIDIA, and Kakao have adopted the same standard, so their watermarked media gets flagged as well. Apple is listed as "coming soon."

Everything else is invisible to it. Microsoft and Meta run their own separate watermark systems, and the labs that dominate AI video culture, Hailuo, Runway, Kling, Pika, Seedance, don't use SynthID at all. If the viral clip in your group chat was made by one of them, this portal will shrug.

What a clean result actually means

A positive result is useful. It tells you the file carries a SynthID watermark, meaning some part of it came out of Google's or a partner's tooling. That's about all it says.

A negative result says almost nothing. It can mean the file is human-made, or it can mean the file was AI-generated by anyone outside the SynthID club. The tool also can't distinguish between "made by AI" and "edited by AI." Google's own FAQ draws the line plainly: this is not a general AI detector. It checks membership, not truth.

The timing helps explain the launch. The EU's AI Act has required machine-readable marking of AI-generated media since August 2, 2026. A public checker gives outsiders a way to verify the watermarking the law demands, while the detector's terms quietly protect it from being probed: reportedly no automated or bulk querying, uploads deleted once results return, and only a 24-hour record kept.

How to use it for real

If you need to check a suspicious file, go to synthid.com, sign in, and upload the original file you received, not a screenshot or a screen recording of it. It works on images, video, and audio. Treat a hit as a lead, not a verdict: confirm what tool made it and whether it was edited afterward. Treat a miss as uninformative, and never cite it as proof a file is real.

Sources

  1. [1] MacRumors, "Google Opens SynthID Detector to Public" (Oct 7, 2026)Read source
  2. [2] The Decoder, "Google says 180 billion images and videos now carry SynthID watermarks as detector goes public"Read source
  3. [3] Unite.AI, "Google Opens SynthID Detector Globally With Partner AI Content Checks"Read source
  4. [4] LAVX News, "Google Opens SynthID Detector to Public"Read source
  5. [5] Superpower Daily, "Google Opens Its AI Watermark Checker Globally With Support Beyond Gemini"Read source
  6. [6] SQ Magazine, "Google SynthID Detector Public AI Media Checks"Read source