SynthID Detector
VerifiedGoogle's free portal for checking media for SynthID AI watermarks
Product preview
SynthID Detector
What it is
SynthID Detector is a web portal from Google DeepMind for checking whether a piece of media was created with AI tools that embed SynthID, Google's invisible watermarking technology. Users sign in, upload an image, video or audio file, and the portal reports whether a watermark is present. Launched in 2025 as a limited tool for journalists and researchers, it was opened to the public in October 2026, according to news reports, and now also recognizes watermarks from partners including OpenAI, NVIDIA and Kakao. It is free with daily usage limits. A negative result does not prove content is human made, since many generators do not use SynthID.
Key features
- Watermark checks for images, video and audio
- Recognizes marks from several AI vendors
- Simple upload-and-scan web portal
- Backed by Google DeepMind research
What we liked
- + Free to use
- + Detects an embedded signal rather than guessing from style
- + Covers several media types
Worth noting
- · Only finds watermarks from participating vendors
- · Daily limits and no bulk checking
Editor's notes
More context from AI BirdView
Many AI detectors try to guess from patterns whether content was machine made, which can produce shaky results. SynthID Detector works differently: it looks for an invisible watermark that participating generators embed at creation time, so a positive match is a much stronger signal. Google built it around its own models and, according to news coverage, has extended it to content from OpenAI, NVIDIA and Kakao.
- Checks images, video and audio files for the SynthID watermark
- Part of Google DeepMind's broader work on responsible AI and provenance
- Recognizes watermarks from multiple AI companies, not only Google
- Free web access for anyone willing to sign in
- Usage rules limit daily checks and prohibit automated bulk scanning
SynthID Detector is best for journalists, fact-checkers, educators and curious users who want to verify whether a suspicious image or clip came from a major AI generator. A practical tip: treat a positive result as strong evidence but a negative result as inconclusive, and pair the check with reverse image search and source verification before drawing conclusions.
Reader reviews
From people who've used SynthID Detector. Emails are never shown publicly.
No reviews yet. Be the first to share your experience.
Alternatives to SynthID Detector
Other tools we like in this space.
Copyleaks
Detect AI writing and plagiarism with forensic accuracy
Originality.ai
The publisher's AI-content and plagiarism watchdog
GPTZero
The AI detector that started in a dorm room