What a C2PA remover changes
Content Credentials can describe where a file came from and how it was edited. They are provenance records, not an automatic declaration that everything in the file was made by AI. Removing embedded credentials removes useful history along with any AI-related assertions they contain.
This tool removes JPEG APP11/JUMBF blocks, recognized PNG credential chunks, WebP C2PA/JUMBF chunks, and recognized C2PA UUID boxes in supported MP4/MOV files. It also removes other metadata under the selected mode; it is not a selective editor for a single signed assertion.
Keep a master, share a copy
- Save your original file with its credentials intact. You may need it later to demonstrate authorship or editing history.
- Inspect the final export in the workspace and review the planned changes.
- Clean the copy, then review the matching media hashes and inspection results.
- Use the copy where appropriate, keeping any required AI disclosure with the content.
Embedded credentials and remote provenance
Removing a local metadata block does not remove a remote manifest, a previously uploaded copy, or a record another service already holds. It also does not invalidate a platform’s own decision about an existing post.
The inspection lists recognized containers. It does not check the signer, validate the trust chain, or certify the authenticity of a file. For that purpose, use a dedicated Content Credentials verifier.
Common questions
Does removing C2PA change image quality?
The cleaner edits supported containers without re-encoding the image. It compares the encoded media payload before and after processing.
Does C2PA always mean AI-generated?
No. Provenance can document camera capture, edits, AI use or other steps. A container alone does not establish which applies.
Can this remove SynthID too?
No. SynthID is an embedded watermark, outside the metadata containers this tool edits.
Sources & further reading
Official references for the platform or format behavior discussed above.
C2PA technical specification