How to check an image, video or audio clip for Google's SynthID watermark

Google said on Wednesday 7 October 2026 that it is opening its SynthID Detector to everyone, "with the tool available globally in English starting today". The portal checks a file for SynthID, the invisible watermark Google and some partners add to AI-generated media. Here is how to use it and what a result means.

What the detector looks for

Google says SynthID uses "imperceptible watermarks across images, video, and audio". According to its announcement, anyone can now "check if an image, video, or audio file was made with AI from Google or our partners, including OpenAI, NVIDIA, Kakao, and soon, Apple."

Read this before you check anything

The detector finds one specific watermark. The portal's own home page says: "This is not a general AI detector. We can only detect media from the above companies, who have adopted SynthID technology."

Its FAQ explains what a negative result means. "Watermark not detected" means "No confident signal was found, so it was not likely to have been edited or generated by a SynthID-enabled AI model. Although this does not totally rule out AI generation or manipulation." The FAQ gives two reasons AI-generated content can come back "not detected":

• "It was created by a model without SynthID (non-partner or pre-adoption)."
• "Heavy modifications degraded the watermark signal."

It also warns that "In very rare cases SynthID can also falsely trigger on non-watermarked content."

So if a check finds no watermark, that does not mean the file is real.

Using the SynthID Detector

Google's post links to the portal at synthid.com. We have not run a check ourselves. What follows comes from the portal's own interface, FAQ and terms text as published on 12 October.

• Sign in and accept the terms. You will be asked to sign in and to accept the portal's terms; the portal says "Please sign in before detection." It offers sign-in with a Google, Apple or ChatGPT account.
• What happens to your file. Google says uploaded files "are processed in real-time. Once the detection results are returned to you, the original media files are immediately deleted." To enforce "daily usage quotas", it keeps a digital signature of each file for 24 hours.
• Choose a media type and upload. The portal lists three types, each with a 100MB maximum:
  • Images: JPEG, PNG, BMP, WebP
  • Audio: MP3, WAV, OGG, AAC, M4A, FLAC
  • Video: MP4, MOV and (except on iPhone and iPad) WebM, marked "(Preview)"
• Read the result. For an image, the FAQ says the detector reports whether a watermark is present. For video and audio, it points to the segments that carry one. For any media type, the portal's code also includes an "Unsure if made with Google AI" result, with reasons such as "Not enough information to detect SynthID."

The portal advises uploading "the highest quality file possible".

On cost and location: neither Google's announcement nor the portal's terms mention a charge, though the terms do refer to daily quotas. Google says the tool is available "globally in English". The post does not list countries.

The check inside the Gemini app

Google's Gemini Apps Help sets out the steps for a computer. You need to be signed in with a personal account or a qualified Workspace account.

1. Go to gemini.google.com.
2. Click Add files, then upload the file.
3. Ask a verification question such as "Is this AI-generated?"

The limits differ from the portal's. You can check one file at a time, up to 100 MB. Videos must be under 90 seconds and audio files under an hour. The help page puts the quota at approximately 10 image, 10 video and 10 audio checks in any 24-hour window.

There is one important difference. Gemini Apps Help says: "While other companies have started to adopt SynthID watermarks, Gemini can currently only recognise content created by Google AI tools." It describes a negative result as meaning the image "wasn't created or edited by Google AI, but it could have been created by other AI systems."

Search and Chrome

In May, Google said you can check an image's origin "by using Search features like Lens, AI Mode and Circle to Search, as well as Gemini in Chrome. Just ask, 'Is this made with AI?' or 'Is this AI generated?'" At the time, Google said the check would come to Chrome "over the coming weeks". Its 7 October post says the built-in checks in Search, the Gemini app and Chrome "now regularly handle over 1 million requests daily". We did not find a Google help page with step-by-step instructions for Search or Chrome.

For a real-world case, see how Full Fact said a viral video of a London protest carried a SynthID watermark.

Why it matters

On our reading, the detector is most useful when it finds a watermark. That suggests the file was likely made or edited with a SynthID-enabled model from Google or a partner, though the FAQ notes rare false triggers. When it finds nothing, it tells you much less. Google's own FAQ recommends "gathering additional evidence". Its Gemini help page suggests some: look for visual errors such as garbled text, run a reverse image search, and check a file's metadata.