On October 7, 2026, Google made SynthID Detector available to everyone worldwide, not just the waitlisted journalists and researchers it launched to at Google I/O on May 20, 2025 — and the portal now also flags AI watermarks from OpenAI, NVIDIA, and Kakao, with Apple joining soon.
2026-10-07 · by Moe Ameen
On October 7, 2026, Google DeepMind opened SynthID Detector — its portal for checking whether an image, video, or audio file carries a SynthID watermark, the imperceptible, machine-readable mark Google embeds in content made with its own AI models — to everyone, worldwide, in English. You upload a file and the portal reports whether a watermark is present and highlights the parts most likely to carry it, rather than returning a blunt AI-or-not verdict. The tool first launched at Google I/O on May 20, 2025 as a waitlist-gated portal for journalists, media professionals, and researchers; this release removes that gate.
The expanded detector isn't limited to Google's own models — Imagen for images, Veo for video, Lyria for audio, and the Gemini models. Google says OpenAI, NVIDIA, and Kakao have also adopted SynthID, so the portal can now flag content watermarked by those companies' tools too, with Apple named as a coming-soon partner. Google now puts the total watermarked at over 180 billion images and videos plus 240,000 years of audio since SynthID launched in 2023 — up sharply from the "over 10 billion pieces" figure it cited when the Detector debuted in May 2025. Google also says built-in SynthID checks inside Search, the Gemini app, and Chrome handle more than 1 million requests a day.
One scope note: at the 2025 launch Google said text watermark detection would roll out within weeks, but Google's current description of the public portal covers image, video, and audio files only — treat text support as unconfirmed until Google's own pages say otherwise. The hard limit that hasn't changed: SynthID Detector only recognizes the SynthID watermark. It cannot tell you something is AI-generated if it was made with a tool that doesn't embed SynthID, and a "no watermark found" result is not proof a human made the file — it verifies one specific provenance signal, not authorship in general. Google has also open-sourced its SynthID text watermarking so developers can build with it, and partnered with GetReal Security on verification.
This is a provenance-and-disclosure story, and it maps cleanly onto how [Kompozy](/) works. Some Kompozy formats lean on Google models — [Persona Photos](/glossary/output-buckets) and Persona Infographics use Gemini face-lock imaging — so those outputs can carry a SynthID watermark and are exactly the kind of asset the Detector is built to flag. The practical move is not to fight that; it's to disclose cleanly and keep the work unmistakably yours. Kompozy's per-post review pipeline is where you do it: before anything ships, you approve each piece and can set the caption, so adding an AI-made label to the right posts is a workflow step, not an afterthought, and [Autopilot](/glossary/autopilot) carries that same caption across the eight social platforms plus blog and email in one pass instead of eight manual edits.
The deeper point is that a watermark says "a machine was involved," not "this is generic." The durable edge is a recognizable voice and a genuine point of view, and that is what the [Persona Brief](/glossary/persona-brief) enforces on every Text Post, Blog Article, Newsletter, and script Kompozy generates. And a large share of what Kompozy makes — captioned [Persona Shorts](/glossary/persona-shorts), brand-exact [Carousel Posts](/glossary/hyperframes), Quote Graphics — is finished, designed content rather than a raw model dump, so it reads as deliberately produced even when a scan confirms AI was in the loop. Disclose the provenance, own the voice, and the watermark stops being a liability.
It is a verification portal from Google DeepMind that checks whether an image, video, or audio file carries a SynthID watermark — the invisible, machine-readable mark Google embeds in content made with its AI models — and highlights the parts most likely to be watermarked. It launched at Google I/O on May 20, 2025 behind a waitlist for journalists, media professionals, and researchers, and Google opened it to everyone worldwide on October 7, 2026.
Content watermarked by Google's own models — Imagen for images, Veo for video, Lyria for audio, and the Gemini models — plus, as of the October 2026 expansion, tools from OpenAI, NVIDIA, and Kakao, with Apple named as a coming-soon partner. It cannot flag AI content made with a tool that doesn't embed SynthID.
No. A clean result only means SynthID couldn't find its watermark. The file could have been made with an AI tool that doesn't use SynthID, so an absent mark is not proof of human authorship — it is the absence of one specific provenance signal, nothing more.
Generally not with routine edits. Google designed SynthID to survive common transformations like cropping, compression, screenshots, and re-saving, so the mark tends to persist where fragile metadata would be stripped. The practical response for creators is to disclose AI use rather than try to defeat the watermark.