AI Metadata Cleaner

## AI Metadata Cleaner — Free In-Browser Metadata and AI Provenance Cleaner

AI Metadata Cleaner (https://aimetadatacleaner.pro/) is a free, client-side privacy utility for people who need to inspect and strip hidden file metadata before they publish images, video, PDFs, or camera RAW files. It runs locally in the browser through WebAssembly. There is no account, no signup, and no file upload to a server. The price is 0: it is 100% free.

The short version is simple. AI Metadata Cleaner inspects and strips EXIF, GPS, XMP, IPTC, C2PA Content Credentials, and AI generator tags (Stable Diffusion, Midjourney, GPT-Image-2, Veo 3) locally in your browser via WebAssembly, with zero file uploads.

### Core capabilities

The tool reads container metadata and then removes it on the device. It strips camera EXIF, GPS location tags, XMP and IPTC blocks, Stable Diffusion and ComfyUI generation prompt chunks, and C2PA/JUMBF container markers. A read-only inspector at https://aimetadatacleaner.pro/detect shows provenance markers without changing the file. Batch mode handles up to 20 files and exports a ZIP in one click. An optional HTML5 Canvas re-encode is available when a fresh pixel encode is wanted for still images.

### Supported file formats

Images: JPEG, PNG, WebP, AVIF, and HEIC. Video: MP4, MOV, and WebM, processed with FFmpeg.wasm. Documents: PDF, processed with pdf-lib, including author, XMP, and document properties. Camera RAW: DNG, CR2, NEF, and ARW through LibRaw WebAssembly. The same local session can cover stills, motion, documents, and RAW without sending the bytes off the machine.

### Honest boundary

AI Metadata Cleaner strips container-level metadata and C2PA/JUMBF markers. Pixel-domain SynthID watermarks are detected and explained, not erased. Invisible frequency-domain watermarks such as Google SynthID are not removed by container cleaning. The tool also does not guarantee how third-party social platforms apply their own AI labels after you upload a cleaned file.

### What C2PA Content Credentials are

Content Credentials are a provenance manifest attached to a file so a viewer can see claims about how the asset was made or edited. In practice those claims travel in a C2PA manifest stored with JUMBF container markers. Stripping that container data is different from repainting pixels. A longer walkthrough of the manifest model is here: [What Is C2PA Content Credentials & How Digital Provenance Manifests Work](https://aimetadatacleaner.pro/guide/what-is-c2pa).

### Step-by-step workflow

1. Open https://aimetadatacleaner.pro/ in a current browser. No signup is required.

2. Optional: open https://aimetadatacleaner.pro/detect and inspect EXIF, GPS, XMP, and C2PA markers in read-only mode.

3. Drop a JPEG, PNG, WebP, AVIF, HEIC, MP4, MOV, WebM, PDF, or RAW file (DNG, CR2, NEF, or ARW). The file stays local.

4. Review the detected EXIF, GPS, XMP, IPTC, C2PA Content Credentials, and AI generator tags.

5. Strip the container metadata. Use Canvas re-encode only if you want a new still-image encode.

6. For a set of files, run the batch cleaner (up to 20) and download the ZIP.

### FAQ

**Is AI Metadata Cleaner free?** Yes. It is 100% free, with no signup, and it runs locally in the browser.

**Does it upload my files?** No. Processing is in-browser via WebAssembly. Files are not uploaded.

**Which AI tags can it see?** It inspects generator tags associated with Stable Diffusion, Midjourney, GPT-Image-2, and Veo 3, plus Stable Diffusion and ComfyUI prompt chunks stored in the file.

**Will this remove a SynthID watermark?** No. SynthID-style pixel watermarks are detected and explained. Container metadata is what gets stripped.

**Can I clean more than one file?** Yes. Batch processing covers up to 20 files with one-click ZIP export.