Will my image lose quality?
No. JPG, PNG and WebP images are cleaned without recompression, so image quality stays the same.
Remove C2PA & AI metadata before you share
See hidden details in your file, remove supported metadata, and save a clean copy.
Free · No account · Your file stays on your device
Removes metadata, not visible or invisible watermarks.
Updated · Maintained by TuliAI · How we test
Before you clean your file
Is my file uploaded?
No. Inspection and cleaning happen in your browser. Your file is not sent to our server, and you do not need an account.
Does it remove C2PA and SynthID?
Supported C2PA Content Credentials can be removed. SynthID and visible watermarks stay in the image.
What can this tool change?
| In your file | What changes | Evidence and limits |
|---|---|---|
| JPG, PNG, WebP metadata | Supported EXIF, GPS, prompts and workflow tags are removed. Image pixels stay unchanged. | The cleaned copy is inspected again. |
| C2PA / Content Credentials | Supported provenance containers are removed. | Container detection does not verify a signature or prove AI authorship. C2PA specification |
| PDF and DOCX metadata | Supported properties are removed. DOCX comments and review history are removed; tracked changes are resolved. Keep the original. | A software name alone does not establish AI use. |
| Visible / invisible watermarks | Remain in the image. This tool does not detect or remove them. | Use a compatible provider verifier. SynthID documentation |
Understand AI metadata and its limits
Check a ChatGPT, Firefly, or other AI image, or a DOCX file, for C2PA Content Credentials and generator tags, then download a clean copy. Invisible watermarks such as SynthID are part of the pixels and stay. Processing stays in your browser.
What is a C2PA remover?
A C2PA remover, also called an AI metadata remover, is a tool that deletes the Content Credentials (the C2PA manifest), prompts, and other provenance metadata stored alongside an image’s pixels. It changes the file’s hidden data, not the picture.
BeforeShare does this in your browser. It removes supported C2PA, EXIF, XMP, IPTC, and PNG text fields from JPG, PNG, and WebP without re-encoding, so image quality is unchanged. Separately, it cleans author and creator properties from PDF and DOCX files. It cannot remove SynthID or other watermarks embedded in the pixels, and a cleaned file does not prove an image was made without AI.
Remove AI metadata from a photo
Choose a JPG, PNG, or WebP to inspect supported EXIF, C2PA, prompt, model, and generator fields, then download a cleaned copy without re-encoding the image.
AI watermark and C2PA checker
BeforeShare detects supported C2PA containers, IPTC AI declarations, and readable generator hints locally. It does not detect or remove SynthID or other content-level watermarks; use the relevant provider verifier when you need to check those signals.
A clean result is not proof
The absence of an explicit marker cannot prove that a file was not AI-generated. Metadata can be absent, removed earlier, or stored in an unsupported location.
Your file stays on your device
Inspection and cleaning run locally in your browser. No account is required and the file content is not sent to our server.
Remove AI metadata from images, PDFs, and Word files
Clean up AI-generated files before you share them. Choose the task that matches your file; each workflow runs locally in your browser with no upload or account.
AI images from ChatGPT, Gemini, Midjourney, or Stable Diffusion
Drop a JPG, PNG, or WebP into the tool above. Supported C2PA Content Credentials, prompts, seeds, model names, and generator tags are listed, then removed without re-encoding the image.
Use the AI metadata remover aboveAI-made PDFs and Word documents
PDF and DOCX files made with AI assistants or code tools can name the software that built them, such as python-docx in a Word file’s creator or description, or a PDF library in the Producer field, along with author and date properties. BeforeShare shows those fields and removes supported document metadata. It does not redact visible text.
Clean a Word documentClean a PDFCheck a file for AI metadata first
Need an AI metadata checker rather than a cleaner? The viewer lists supported C2PA evidence, IPTC AI declarations, prompts, and creator-tool fields without changing the file. Finding nothing cannot prove that a person made the file.
Open the metadata viewerMany AI files at once
Clean up to 20 images, PDFs, or Word files in one batch, download them as a single ZIP, and keep a local report with the SHA-256 fingerprint of each clean copy.
Open batch cleanupEvidence and verification
BeforeShare reports deterministic file evidence. Use these primary sources to interpret what a marker can — and cannot — establish.
C2PA technical specification
Defines the provenance container and trust model. BeforeShare detects supported container evidence but does not validate its signature or trust chain.
Read the C2PA specificationContent-level watermarks
SynthID and similar provider signals are not ordinary metadata. They require the provider’s compatible verifier.
Read the SynthID documentationWhich AI tools use watermarks?
Compare current provider claims, open provenance, invisible signals, visible labels, and the limits of a negative result.
Read the evidence briefClean metadata from real AI platform exports
Platform behavior changes by product, model, file type, export path, and date. Inspect the exact downloaded file; these provider-specific notes explain what BeforeShare can remove and what can remain.
ChatGPT, DALL·E, Codex, and the OpenAI API
Current supported OpenAI image exports include C2PA metadata and a SynthID watermark, although coverage can vary. BeforeShare can inspect and remove supported C2PA and ordinary metadata without re-encoding the image; it cannot detect or remove the embedded SynthID signal. Keep the original if provenance matters and use OpenAI Verify when you need provider verification.
Provider source: OpenAI provenance signalsAdobe Firefly
Adobe says it automatically applies Content Credentials to assets whose pixels are fully generated with Firefly. BeforeShare can remove the supported credential container from the local JPG, PNG, or WebP copy, but it does not validate the credential and cannot erase any durable copy stored in Adobe’s Content Credentials cloud.
Provider source: Adobe Content Credentials overviewMicrosoft Copilot and Designer
Microsoft documents Content Credentials for AI-generated images from Copilot and Designer. BeforeShare reports supported C2PA container evidence and readable generator hints, then removes the supported container from the cleaned copy. It does not verify the signature, issuer, or trust chain.
Provider source: Microsoft Designer FAQGoogle Gemini, Nano Banana, and Imagen
Google-generated images can use SynthID, a content-level watermark rather than ordinary file metadata. BeforeShare can remove standard metadata containers that happen to be present, but it does not detect or remove SynthID. A clean local metadata result is therefore not a negative SynthID result.
Provider source: Google Imagen documentationStable Diffusion, AUTOMATIC1111, and ComfyUI
AUTOMATIC1111 can store prompts, negative prompts, seeds, samplers, CFG values, and model details in PNG text. ComfyUI can store the prompt and entire workflow JSON, including in animated PNG comf chunks. BeforeShare now shows those fields as AI evidence and removes them from the cleaned copy; save the original first if you need to reproduce the workflow.
Provider source: AUTOMATIC1111 PNG infoProvider source: ComfyUI metadata writerMidjourney, Runway, Leonardo, Ideogram, and other tools
Do not assume a platform always adds—or never adds—a particular field. Export formats and downstream editors change. BeforeShare checks supported C2PA evidence in JPG, PNG, WebP, and DOCX. It separately checks ordinary image metadata and PDF or DOCX document properties, without guessing the source from appearance.
What Is AI Metadata in Files?
Images may contain prompt, model, creator-tool, timestamp, or C2PA-container evidence. PDF and DOCX files may contain general document properties. Generic software metadata alone cannot prove that AI created a file.
C2PA / Content Credentials
C2PA is a provenance standard for media files. BeforeShare reports supported container evidence and readable generator hints; it does not validate signatures, trust chains, or every assertion.
Prompt and model metadata
Some generators and workflows store prompt text, negative prompts, seeds, samplers, model names, checkpoints, LoRA names, or workflow JSON in XMP or PNG text fields.
Visible watermarks
A visible label, logo, signature, or platform badge is part of the image pixels. Metadata cleanup does not remove visible content.
What Changes When You Clean the File
What stays
The visible subject, dimensions, edits, and crops remain. JPEG, PNG, and WebP image data is copied without re-encoding, so the picture and its quality are unchanged; only metadata segments are removed.
What goes
Supported EXIF, XMP, IPTC, PNG text, creator-tool, timestamp, model-setting, and content-credential containers are removed. The receipt re-inspects the output for removable fields the detector supports.
What to review
If the image visibly contains a watermark, signature, prompt text, username, or private detail, remove that separately before sharing.
Supported File Types and What We Check
Images
JPG, PNG, and WebP files can carry EXIF, XMP, IPTC, PNG text chunks, Content Credentials, creator-tool fields, prompts, seeds, and model names.
PDF reports
AI-generated PDFs commonly keep Title, Author, Subject, Creator, Producer, Keywords, timestamps, and sometimes XMP metadata.
Word documents
DOCX outputs can include creator, last modified by, company, revisions, editing time, template, app version, and a C2PA 2.4 manifest at the standard package location.
How to Remove C2PA Metadata from an Image
- Choose the file — Choose JPG, PNG, or WebP for image provenance checks. PDF and DOCX are checked separately for general readable metadata.
- Review AI-related fields — Inspect supported C2PA and Content Credentials evidence in images. Review creator-tool, prompt, model, PDF property, or DOCX fields separately.
- Download and verify the copy — Remove supported metadata, review the post-clean receipt, and save the new copy.
AI Metadata FAQ
How do I remove C2PA metadata from AI generated images?
Drop your AI-generated image into the tool above. BeforeShare reads the file in your browser, reports C2PA content credentials and other readable metadata, then creates a clean copy without supported metadata containers. No upload or software installation is required.
How to remove content credentials from an image?
Content Credentials use C2PA provenance data. For supported JPG, PNG, and WebP images, BeforeShare creates a copy without supported metadata containers. JPEG, PNG, and WebP are all cleaned at the container level without re-encoding.
How do I clean ChatGPT image metadata?
Some ChatGPT or DALL-E image exports may contain C2PA or creator-tool metadata, depending on the product, settings, and export path. Inspect the actual file above, then remove supported containers and review the post-clean receipt.
Can BeforeShare detect SynthID or an invisible AI watermark?
BeforeShare does not detect proprietary content-level watermarks from ordinary metadata. It checks supported C2PA, IPTC, prompt, model, and creator-tool fields locally. Use the relevant provider verifier for SynthID, Amazon, or another provider-specific watermark, and treat a negative result as inconclusive.
Can I remove an AI watermark or Content Credential when disclosure is legally required?
Do not remove or hide a required disclosure, provenance marker, platform label, evidence record, or contractual notice. Rules vary by jurisdiction and use. Keep the original and get legal advice when the EU AI Act, China’s labelling rules, or another obligation may apply.
Does this remove visible AI watermarks from images?
No. BeforeShare removes hidden metadata only — C2PA manifests, content credentials, EXIF, XMP, and IPTC fields. Visible watermarks, logos, signatures, or text baked into the image pixels are part of the picture itself and require image editing to remove.
Does the image get uploaded to a server?
No. BeforeShare processes the file locally in your browser with JavaScript, and the page does not send the file’s contents to our server. Your file stays on your device.
Is there a free AI metadata remover that doesn’t upload my files?
Yes. BeforeShare is free, needs no account, and runs in your browser, so the file is not sent to our server. It removes supported AI metadata from JPG, PNG, and WebP without re-encoding, and cleans general metadata from PDF and DOCX files.
How do I remove AI metadata from a PDF or Word document?
Choose the PDF or DOCX file in the tool above. BeforeShare lists readable document properties such as author, creator, producer, description, and dates. Files generated by code tools can name the library that built them, for example python-docx. It then creates a clean copy without supported document metadata. For DOCX, supported C2PA evidence at the standard package location is removed too.
How can I tell if an image has AI metadata?
Drop it into the checker above or open the metadata viewer. BeforeShare lists supported AI evidence, such as C2PA containers, IPTC digital source type, prompts, seeds, and generator names, before anything is removed. An image with no marker may still be AI-generated, because metadata can be missing or already stripped.
Can I remove AI metadata from multiple files at once?
Yes. The batch tool cleans up to 20 JPG, PNG, WebP, PDF, or DOCX files (100 MB in total) and downloads them as one ZIP with a local verification report.
Does it work on iPhone and Android?
Yes. It runs in modern mobile browsers such as Safari and Chrome. Choose the file from Photos or Files, then save the clean copy; there is nothing to install. HEIC photos are not supported yet, so export them as JPG first.
Can I remove SynthID and C2PA at the same time?
No. BeforeShare removes C2PA Content Credentials and other supported metadata, but SynthID is not metadata: Google embeds it in the image pixels, so no metadata cleaner removes it. After cleaning, a SynthID verifier can still identify the image, and the cleaned copy can still be labeled as AI by platforms that check pixel watermarks.
I edited the image in Canva, Lightroom, or Photoshop. Does it still work?
Yes, for exported JPG, PNG, and WebP files. Lightroom and Photoshop can attach Content Credentials on export when that option is on, and Lightroom Classic always attaches them once generative AI tools were used. Canva does not publicly document C2PA on its exports, so check the file above to see what it actually contains. Canva’s AI terms also ask you not to remove provenance tags from AI-generated content, so check the terms of the tool you used.
Can it remove C2PA from videos (MP4 or MOV)?
Not yet. AI video tools can attach C2PA to MP4 and MOV files, but BeforeShare only cleans JPG, PNG, and WebP images and document files, so a video is rejected instead of being partly cleaned.
How do I remove AI metadata from an image, step by step?
Choose the JPG, PNG, or WebP in the tool above; it opens on your device. Review the fields it finds, such as Content Credentials, prompts, model names, and camera data. Select “Remove metadata & download”, check the receipt, which inspects the new copy again, and save it. The steps are the same for images from ChatGPT, Gemini, Midjourney, Firefly, or Stable Diffusion. Watermarks inside the pixels, such as SynthID, stay.
What is C2PA metadata?
C2PA metadata, shown to people as Content Credentials, is a signed record attached to a file. It can say which app or AI model created it, when, and which edits followed. It is an open standard from the Coalition for Content Provenance and Authenticity, whose members include Adobe, Microsoft, Google, and OpenAI. In an image it is stored in its own block next to the pixel data, so it can be read or removed without changing the picture.
Can C2PA metadata be removed?
Yes. The manifest is stored next to the pixels, not in them, so deleting that block removes the Content Credentials from the file, and a C2PA checker then finds none. Some providers can still recognize the image afterwards: Google’s SynthID is a watermark in the pixels, and Adobe can match some images to credentials kept in its cloud. Keep the original if you may need to show where an image came from.
How do I remove Adobe Firefly Content Credentials?
Download the Firefly image as JPG or PNG and choose it in the tool above. BeforeShare lists the Content Credentials it finds and saves a copy without the supported container, without re-encoding the image. It cannot delete a copy of the credential stored in Adobe’s Content Credentials cloud, so Adobe’s own tools may still match the cleaned image.