FAQ · AI content & watermarks
Only partly — a re-encode disrupts some invisible patterns but leaves C2PA, generator tags and the encoder fingerprint, and robust watermarks can survive.
Partly. A controlled re-encode disrupts many invisible pixel-domain patterns, but robust watermarks can survive it, so re-encoding alone isn’t a guarantee. It also does nothing about the biggest giveaways: C2PA Content Credentials, generator tags and the encoder fingerprint (like FFmpeg’s “Lavc libx264” name and the embedded x264 SEI) that mark the file as non-phone footage. A naive re-encode can even stamp a fresh editor signature that itself flags the file. The dependable approach is to strip the metadata container and encoder fingerprint you can fully remove, then inject an authentic phone encoder and device profile.
Related on Calabi: Does Instagram detect AI images? · How do I remove C2PA metadata? · Does cropping remove a watermark? · How to post AI content on Instagram without getting shadow-banned · Sora — or just open the tool.