Watermarks RemoverFree browser-local tool
Image files can carry location, camera, date, software, descriptive, workflow, and provenance information that is not visible in the picture. This free tool scans supported JPG, PNG, and WebP files locally, groups the records it understands, and offers a rights-preserving privacy preset or a separately confirmed Full Metadata Reset. The privacy preset targets supported GPS, device, timestamp, software, user, and AI-generation blocks while preserving recognized author, copyright, license, attribution, and rights-holder records. Every cleanup mode preserves dimensions, encoded pixels, ICC color profiles, orientation, transparency, and animation. Unsupported or malformed data is reported rather than silently discarded. Metadata cleanup cannot remove visible text or logos, change information already copied elsewhere, or erase signals embedded in image pixels.
We’ll check for metadata, Content Credentials and supported hidden markers before you choose how to clean it.
JPG, PNG and WebP · up to 20 files · 25MB per file
JPG, PNG and WebP · Only upload files you own or are authorized to edit
Supported removal types
The privacy scan organizes supported records by purpose instead of presenting a single destructive “strip all” switch. Location can include GPS coordinates and related positioning fields. Device data can include camera make, model, lens, serial-like identifiers, and capture software. Date information can include original, digitized, or modification timestamps. AI data can include prompts, seeds, models, workflows, and generator labels.
Descriptive XMP, IPTC, EXIF, PNG text, and WebP metadata may contain both private details and valuable rights information. The result screen therefore shows recognized author, copyright, license, attribution, and rights-holder fields in a protected group. C2PA is shown as a distinct provenance category because removing a signed manifest carrier is not equivalent to deleting a simple text tag.
How it works
Inspection, preset choice, rebuilding, and downloading take place in the browser without requiring sign-in.
Choose, drop, or paste an authorized JPG, PNG, or WebP image. The parser checks its signature and container boundaries before reading records.
Review the scan cards for privacy metadata, AI metadata, C2PA, rights information, and unsupported records.
Use the fixed privacy preset, or explicitly confirm Full Metadata Reset when you have permission to remove recognized rights and provenance blocks too.
Build a copy that omits the blocks included by the chosen preset and copies the original encoded media payload without decoding it.
Read the report for removed, preserved, and unsupported items, confirm that pixels were not re-encoded, and download the result.
Metadata vs pixels
EXIF, XMP, IPTC, C2PA, ICC, and format-specific text live in container structures associated with encoded media. The cleaner rebuilds the container around the retained JPEG scan, PNG image data, or WebP image chunks. The implemented default and Full Reset modes therefore keep the visible picture and dimensions unchanged.
Visible overlays are rendered content, while invisible pixel watermarks alter or encode patterns in the pixel signal. Neither disappears when only container metadata is edited. The clean report makes this boundary explicit instead of claiming that a metadata-free copy has no other watermark layer.
What is preserved
Every cleanup mode retains orientation so a portrait does not unexpectedly rotate, ICC color data so color-managed software can reproduce the intended appearance, and transparency or animation structures needed by PNG and WebP. Encoded image and animation payloads are always copied rather than re-encoded.
Author, copyright, license, attribution, and rights-holder records stay protected in the default preset. If removable privacy data shares a block with recognized rights information, the default keeps that block and reports the limitation. Full Metadata Reset may remove a recognized rights or provenance block only after permission is explicitly confirmed; technical image data and pixels still stay preserved.
Limitations
Applications can store metadata in private namespaces, sidecar files, cloud databases, filenames, or container variants this parser does not support. Some records may be compressed, duplicated, malformed, or linked externally. The scan is evidence about this file, not a promise that no personal or provenance information exists anywhere else.
Cleaning does not recall an image already shared, modify platform logs, remove a visible copyright mark, or affect pixel-level signals. Review the cleaned report before publishing. Do not process identity documents, government records, financial materials, certificates, legal evidence, or third-party commercial imagery you are not authorized to edit.
Reproducible example
A generated animated WebP fixture exercises the fixed privacy preset while keeping its media payload independent from any personal photograph.
FAQ
GPS, precise timestamps, device identifiers, and private comments are common privacy choices. Keep rights and license information unless you own it and have a clear reason and authority to remove it.
No. It applies the fixed privacy preset and keeps orientation and protected rights records. Use the EXIF-specific page for a preset limited to supported EXIF privacy data.
The lossless path preserves recognized PNG or WebP animation structures and payloads. If the parser cannot safely preserve an animation, it should refuse that clean rather than silently return a single frame.
No. Unknown namespaces, filenames, external systems, visible content, and pixel signals may still disclose information. The scan covers only supported records in the selected file.
Related tools and guides
Only clean files you own or have permission to edit. Privacy cleanup must not be used to destroy evidence, alter official records, conceal infringement, or remove another creator's rights information.