Google opened its SynthID Detector website to everyone on October 7, after more than a year with early testers only. People can upload a file and see if it carries a SynthID watermark. It works in English and reads about 20 file types. Google says it cannot catch every AI-made file.
New to this? Read it in simple words
- Google’s SynthID Detector website is now open to everyone, for free.
- It checks pictures, video and audio for hidden AI watermarks.
- It only finds media from companies that use SynthID, such as Google and OpenAI.
- A file with no watermark found can still be made by AI.
- Watermark
- A hidden signal inside a file that shows where it came from. People cannot see or hear it.
- SynthID
- Google DeepMind’s system for hiding watermarks in AI-made pictures, video, audio and text.
- C2PA
- A shared standard for labels that record how a file was made and edited.
Who can use it
The SynthID Detector website is now open worldwide, in English. 9to5Google reports that more languages are planned.
Engadget reports that it is free, but users must sign in with a Google, OpenAI or Apple account. Users upload a file, and the site says if it carries a SynthID watermark.
TechCrunch lists 20 supported file types: 11 for images, 3 for video and 6 for audio. They include JPG, PNG, MP4 and MP3.
From Google’s announcement and reports by TechCrunch and Engadget.
Which AI media it can find
The detector finds SynthID watermarks from Google, OpenAI, Nvidia and Kakao. Google says Apple’s content will follow “soon”.
It is not a general AI detector. TechCrunch notes that Microsoft and Meta use other standards, so their AI media will not show up. Engadget reports that the site also reads C2PA Content Credentials, another labelling system.
Google says more than 180 billion images and videos and 240,000 years of audio have been watermarked since 2023. These are Google’s own figures.
The limits
Google and the reports agree that the tool cannot catch every AI-made file. Watermarks can sometimes be removed. The site also does not say which parts of a file were made or edited by AI.
Google has not published how often the detector misses a watermark. In July, the fact-checking site Lead Stories found a bug in the SynthID check inside the Gemini app, which Google said it fixed.
Sources
Every fact in this story comes from the sources below. Open them to check our work.
- 1Primary source · October 7, 2026We’re making it easier to identify AI-generated content globally Google
- 2
- 3
- 4Research · October 7, 2026Google’s SynthID AI content detector expands globally, will work on Apple’s models ‘soon’ 9to5Google
- 5Research · July 13, 2026Warning: Google Gemini SynthID AI Watermark Detector Appears To Mix Up Results In Same Chat Lead Stories
The watermark totals are Google’s own figures, and Google has published no miss rate. No outlet has tested the new website yet.
A free public checker for AI-made media opens worldwide, but it misses media from other makers and can be fooled. How we rate the mood