SynthID & AI Dating Photos
SynthID is Google's invisible watermark — baked into the pixels, not the file's metadata. Here's what it actually means for an AI dating photo, why metadata strippers can't touch it, and the one reliable way to avoid it.
No public tool can read SynthID itself — but you can check the provenance metadata that often travels with it.
What is SynthID?
SynthID is an invisible watermark developed by Google DeepMind. When a supported Google model generates an image — Imagen, Gemini's image tools, Veo for video — it makes tiny, imperceptible adjustments to the pixels in a pattern a matching detector can recognize. You can't see it. A compatible detector can.
The important part is whereit lives. C2PA Content Credentials and EXIF tags sit in the file's header, alongside the pixels — readable, and removable by re-encoding. SynthID is woven into the pixels and is designed to survive cropping, resizing, and recompression. That makes it a fundamentally different problem from metadata.
Why it matters for dating photos
For most of SynthID's life it was niche. That is changing. Other AI providers have started adopting SynthID-style watermarks, and Google has begun surfacing “made with AI” signals directly in Chrome and Search. As detection spreads into browsers and platforms, an image that carries a durable AI watermark becomes easier to flag as AI-generated — long after you posted it.
The practical takeaway: if your photos came from a tool that embeds SynthID, that signal can ride along even after you've stripped every visible trace of metadata. For a dating profile you want to last, that's a quiet liability worth understanding before it matters.
Can you detect SynthID?
Not with a general public tool. Google's own SynthID Detector is access-gated, and no reliable third-party detector exists for the pixel watermark. What you can check is provenance metadata — C2PA, EXIF, and AI tags — with a free metadata checker. If those show a Google AI origin, SynthID is likely present too.
Can you remove SynthID?
Not reliably. It's built to persist through ordinary edits, so a metadata strip does nothing to it, and the aggressive editing that might disrupt it usually ruins the photo first. Any tool promising to perfectly erase an invisible pixel watermark with no quality loss is overstating what's possible. See what's actually removable.
How MatchMaxing avoids SynthID
You can't reliably scrub a pixel watermark after the fact — so the only dependable answer is to never introduce one. That's the approach behind our photos:
- Generated on open models that don't embed SynthID. There's no pixel watermark to begin with, because the models we use don't add one.
- Re-encoded clean. Every file is re-encoded as a standard JPEG, so no C2PA Content Credentials or EXIF AI tags ship with it either.
- Built from your own selfies. The result looks like you — a normal photo file, not a flagged one.
Want the full picture on AI fingerprints? Read how SynthID works, SynthID vs C2PA, and our guide to AI photo watermarks for dating apps.
Frequently asked questions
What is SynthID?
SynthID is an invisible watermark from Google DeepMind that is woven into the pixels of an AI-generated image. Unlike metadata in the file header, it is designed to survive cropping, resizing, and recompression.
Can you check an image for SynthID?
Not with a general public tool. Google's SynthID Detector is access-gated, and there is no third-party detector that reliably reads the pixel watermark. The signal you can practically check is provenance metadata such as C2PA and EXIF tags.
Can you remove SynthID?
Not reliably. It is built to persist through ordinary edits, so stripping metadata does not remove it, and aggressive editing tends to wreck the photo before it removes the watermark. The dependable path is to start with an image that never had one.
Do MatchMaxing photos contain SynthID?
No. We generate on open models that do not embed SynthID, so there is nothing to strip — and we re-encode every file to drop C2PA and EXIF metadata as well.