Help with your files
Use AiMetaOff to inspect files, choose what to remove, and save a separate copy. If a file is rejected or a result looks unexpected, start with Troubleshooting below.
Getting started
Open the AiMetaOff workspace or Inspect File.
In the drop area, select Choose files or drag and drop files. You can add several files at once.
Wait for each file to finish inspection. AiMetaOff reads files one at a time. Cleanup starts when you select Clean selected.
Inspecting files
Start with the summary on the file card. Open View technical details for supported metadata findings, inspection limits, and any available Content Credentials results for the original file.
Found means inspection detected a supported record.
Not found means the inspection found no record in the structures it checked.
Inspection unsupported or Unable to inspect means the report cannot determine whether the record is present.
Choosing metadata to remove
In a fresh workspace, all six Cleanup scope boxes are checked: EXIF, XMP, IPTC, Text, C2PA, and Info. Clear the boxes for categories you want to keep. Your choices apply to the selected files, and each format supports its own set of categories.
AiMetaOff removes whole supported categories. For example, removing Text or XMP can remove captions along with prompts. The AI Metadata Remover uses these same controls. If you only want to remove supported embedded Content Credentials, leave only C2PA checked. Open C2PA Remover.
Files are selected when added. Use the checkbox next to each filename to include or exclude that file. Select all eligible toggles the eligible file rows. To clear the cleanup categories, uncheck their boxes individually.
Cleaning and downloading
Select at least one eligible file and one category, then choose Clean selected. AiMetaOff checks the output before offering a download and leaves your original file as it was.
Open View technical details to compare Input and Output and review what was removed, retained, or outside the inspection scope. Supported cleanup verified means the selected supported cleanup passed the output check. No selected data to remove means the tool found nothing it could remove in the selected scope.
Choose Download new copy on the file card, or Download completed ZIP to save eligible copies together. The ZIP uses all eligible generated copies still in the workspace, regardless of which rows are currently checked. Limited copies have an individual download marked Limited cleanup.
Before sharing, open the downloaded copy and check its appearance or readability. For a second inspection, follow the metadata removal verification guide.
Supported formats and limits
Format | What the tool can inspect and clean |
|---|---|
JPEG (JPG) | Supported EXIF, XMP, text comments, IPTC-related blocks and embedded C2PA in single-image files. |
PNG | Supported EXIF, XMP, text and embedded C2PA in static images. |
WebP | Supported EXIF, XMP and embedded C2PA in static images. |
AVIF | Supported EXIF, XMP and recognized embedded C2PA in certain single-image files. Grids, sequences and extra image items such as alpha, depth and thumbnails are outside the supported subset. |
HEIC (limited) | Supported EXIF, XMP and recognized embedded C2PA in certain single-image files. Preview availability depends on your browser. Codec-internal metadata can remain. |
Document properties (Info) and document-level XMP within supported PDFs. Supported signed PDFs are inspection-only. Encrypted files and PDFs with compressed object streams are rejected. |
The tool checks file contents and structure. GIF, animated images, and unsupported multi-image variants are rejected. Even a file within the limits below can be too complex or use an unsupported structure.
Queue: Up to 30 files, with a maximum of 15,000,000 bytes (15 MB) per input file.
Image and PDF: Up to 40 megapixels per image and 200 pages per PDF.
Processing: 20-second worker limit for file processing, including inspection or cleanup, and for ZIP creation. Batch time depends on the files and the browser.
Outputs: Up to 100,000,000 bytes (100 MB) of generated copies in the workspace. The completed ZIP also has a 100 MB limit on eligible output bytes before compression.
Troubleshooting
A file cannot be opened
Check Supported variants & limits in the drop area and read the message on the file card. “This file format or variant is not supported” can mean the file uses an unsupported structure even if its extension is listed. If the file is damaged, try a fresh export from the original application. For encrypted PDFs, use an authorized unencrypted source copy.
Metadata is missing from the report
Open View technical details and check the status for that category. The file may lack that record, use a metadata structure the tool cannot read, or rely on information stored separately by an app or website. Inspect the original export if you have it. A compressed PDF XMP stream can be detected without displaying its values.
Some metadata remains after cleaning
Check Cleanup scope and the output's Retained and Inspection limits sections. If you left a category unchecked, select it and run Clean selected again. For supported image cleanup, AiMetaOff keeps the compressed image data without re-encoding. It keeps supported color information and any orientation data needed for display. Some format-specific or vendor records can remain outside the supported cleanup scope. If you are checking GPS, use the guide to removing photo location data.
C2PA credentials cannot be verified
In View technical details, look for Verification results when the file has a readable manifest.
If no embedded credentials were found, open the original export if available. An external credentials message means the external record was unavailable to the local check.
If Claim signature or File content binding shows failed, review the reported check before relying on the credentials. Signature and file binding are separate checks.
An unsupported signature algorithm or unreadable structure can leave verification unavailable.
Signer trust remains unverified locally. The tool does not check current revocation or external identity.
A PDF cannot be cleaned
If Inspection limits shows Digitally signed PDF: cleaning is disabled to protect signatures, you can still inspect supported properties. Ask for an unsigned source copy if you need to clean it. The tool also rejects encrypted PDFs, compressed object streams, and some invalid or complex files. If you can re-export from the source document, try a supported unsigned and unencrypted PDF. PDF cleanup addresses document properties and document-level XMP. Visible text, forms, attachments, annotations and embedded-image metadata require separate review. See the PDF metadata guide or open PDF Metadata Remover.
Processing takes too long
Check the card for Reading, Inspecting, Cleaning, or Verifying output. Files wait their turn. If you see “Processing stopped after 20 seconds. Try a simpler file,” try a smaller or simpler source export. For a slow batch, try fewer files at a time. To stop working on a particular file, use its Remove file control. To retry inspection, add the file again.
A download is missing or the ZIP fails
Wait for processing to finish. If a file has no download link, check its status and any error on the card. The ZIP includes verified and unchanged copies still in the workspace. Limited cleanup results are available individually when a link is shown. If ZIP creation fails, use Download new copy on the available cards. If the workspace reaches the 100 MB output limit, download the copies you want, remove their rows to free space, and process the remaining files again.
Files disappear after reload
During supported navigation in the same tab, Back to your files returns to the workspace. A full reload or closing the tab clears the queue, results, and cleanup choices. Add your original files again to resume. Download any copies you need before refreshing or closing the tab.
A social media AI label remains
Platforms can use disclosures, watermarks, or other signals beyond ordinary embedded metadata. Cleaning a local file leaves an existing post as it was. See the guide to AI labels and image provenance.
Privacy and processing
Your file contents stay in browser memory and are processed on your device. Loading the site and its processing components can require ordinary website requests. AiMetaOff does not follow links embedded in the file. Read the privacy policy.
Where to learn more
Visit the AiMetaOff FAQ for general questions about metadata, privacy, and what cleanup can do.