Guide

What Is SynthID?

SynthID is an imperceptible watermark technology from Google DeepMind. When supported AI systems generate or edit media, it can embed a signal into images, video, audio — and in some cases text — that a detector can later look for.

Why watermarks exist

As generative AI spreads, publishers, platforms, and brands need a way to flag AI-touched media without ruining the file for viewers. SynthID aims to add a signal that is hard for people to notice, yet still recoverable by a matching detector under many common edits (crop, mild compression, etc.). Coverage and robustness depend on the partner system and how the file was processed afterward.

What SynthID is — and is not

Who participates (high level)

Google has positioned SynthID Detector as a public portal for checking watermarked media from participating partners. Public reporting has mentioned ecosystems involving Google and other companies (for example OpenAI, NVIDIA, Kakao, with Apple noted as coming in some announcements). Partner lists and formats change — always confirm on the official site before you rely on a check for publication.

Official detector: https://synthid.com

How this helps journalists and marketers

Use SynthID when you already suspect a file may come from a supported pipeline, or when your workflow asks for an AI-watermark check before publish. Combine it with provenance questions, source interviews, and editorial judgment. For a practical walkthrough, see How to use SynthID Detector. To understand the difference from generic AI detectors, read SynthID vs AI detectors.

Ready to check a file? Free spot checks run on Google’s official tool. Need a batch workflow or checklist? See pricing on this site.

Disclaimer

checksynthid.com is an independent educational and services site. We are not Google, DeepMind, or any SynthID partner. We do not operate Google’s detection engine. Free checks should be performed at synthid.com.