Overview
AI Metadata Cleaner is a small Windows application that decodes selected images and exports them as fresh PNG files. It does not copy EXIF, XMP, IPTC, color-profile, C2PA/Content Credentials, or similar embedded blocks into the new file. Visible image content, watermarks, logos, and signatures are not changed.
Problem
Image files can carry embedded metadata and provenance information that may not be intended for a redistributed copy. Manual cleanup varies by file format and does not always provide a clear workflow that preserves the visible image.
Solution
The application reads image pixels, creates a new PNG instead of copying the source file structure, and validates the output. Users can select images, choose an output folder, track progress, and save results under separate filenames.
Key features
- Image selection for PNG, JPEG, WebP, BMP, and TIFF inputs.
- Informative scanning for EXIF, XMP, IPTC, color-profile, and C2PA/Content Credentials markers.
- Pixel-preserving cleanup flow that exports a fresh PNG.
- Multiple-file processing, output-folder selection, progress, and error summaries.
- English/Türkçe interface and Windows packaging with PyInstaller.
Technical approach
The tool uses Python, Tkinter, and Pillow, with PNG chunk inspection and output validation. PyInstaller creates the Windows package and GitHub Actions supports the Windows build workflow.
My role
Product concept, bilingual interface, image-processing workflow, PNG output validation, and Windows packaging.
Usage note
Removing C2PA/Content Credentials can invalidate the file’s provenance record. The tool should be used only with files the user owns or is authorized to process.