About

Why this exists

Search for "AI watermark remover" and you will find a lot of tools promising things nobody can deliver — guaranteed SynthID removal, content that is "100% undetectable", detectors for watermarks that cannot be detected client-side. Most of them upload your file to a server on the way.

NoWatermark is the opposite proposition. It does a smaller set of things, does them properly, and is explicit about where the boundary is.

Two rules

Never claim a watermark has been removed unless we can verify it. Our cleaners scan their own output and report the difference between two independent scans. A signal is marked removed because a fresh scan cannot find it — never because a cleaner said so. Signals we cannot measure are marked "Unable to verify" and never appear in a removed list.

Your file never leaves your device. There is no upload endpoint, no database, no storage. Load a page, disconnect from the internet, and the tools still work.

Provider-neutral

We cover OpenAI, Google, Anthropic, Adobe and the open standards equally, and we have no stake in which of them you use. Where a provider does something well, we say so; where a technology defeats us, we say that too.

How it is built

A static site with no backend. The scanners and cleaners are hand-written container parsers for JPEG, PNG and WebP, running in a Web Worker on your device. Cleaning is lossless by construction: compressed image data is copied byte-for-byte and never re-encoded.

The methodology page sets out the full capability matrix, including everything we cannot do.

Contact

Corrections are genuinely welcome — particularly if you think we have described a technology inaccurately. hello@nowatermark.fyi.