Stable Diffusion Image Watermark Remover

Stable Diffusion (depending on the UI) writes prompt, model and seed into PNG text chunks and may apply Stability's invisible-watermark library. We strip the PNG metadata.

⚠️ Stable Diffusion does not use C2PA. Our tool will re-encode the image and strip standard metadata (EXIF, IPTC, PNG text chunks). Pixel-level watermarks like visible logos or SynthID are not removed.

Drop your ChatGPT images here

JPEG · PNG · WebP — up to 50 images — Max 50 MB each

or click to browse

All processing happens locally in your browser. Your images never leave your device.

About the Stable Diffusion / Stability AI watermark

Stable Diffusion outputs through Automatic1111, ComfyUI and similar UIs embed your prompt, negative prompt, seed, sampler and model hash as PNG tEXt chunks — anyone with a metadata viewer can read your entire workflow. The official Stability release also applies their invisible-watermark library by default. Our canvas re-encoder removes all PNG text metadata. The invisible-watermark layer is pixel-level and partially survives re-encoding.

Looking for the ChatGPT version?

Our primary tool is built for ChatGPT. See the ChatGPT Image Watermark Remover or the full comparison.

Ready to remove your ChatGPT watermark?

100% free · no login · images never leave your browser

Open the Free Tool →