Location
GPS fields can expose precise coordinates, altitude, and direction. A photo taken at home, work, or while traveling may reveal more than the visible scene.
GPSLatitude / GPSLongitude / GPSAltitude / GPSImgDirection
AI workflow / EXIF / GPS / PDF XMP
Choose a supported image or PDF. We detect the file type locally, open the matching metadata workflow, and verify what changed without uploading the original.
Local file detection
Ready to detectFree · no signupThe browser checks the file signature first. Supported images stay in the AI metadata cleaner; PDFs open the document-property cleaner.
We check its contents locally instead of trusting the extension.
JPEG, PNG, static WebP, PDFChoose one file first · images can add up to 5
18 image + 4 PDF fixtures · cross-browser tested
View test evidenceAI metadata removal should be observable, not a black box. The workbench shows each finding and decision while keeping the original file unchanged.
Read supported prompts, generation settings, EXIF, GPS, and provenance markers before changing anything.
Separate file-level AI workflow data from location, identity, device, rights, display, and provenance fields.
Re-encode the visible image locally and discard metadata that is not needed for display.
Inspect the new file again, list what was removed, and mark anything the scanner cannot verify.
A finished image can still carry prompts, model and workflow settings, provenance records, location, device, author, and editing history.
GPS fields can expose precise coordinates, altitude, and direction. A photo taken at home, work, or while traveling may reveal more than the visible scene.
GPSLatitude / GPSLongitude / GPSAltitude / GPSImgDirection
Author names, camera models, device identifiers, and serial numbers can connect an image to a person or a specific piece of equipment.
Artist / Creator / Make / Model / SerialNumber
Capture, digitized, and modification dates can show when an image was created or moved through an editing workflow.
DateTimeOriginal / CreateDate / ModifyDate
AI-generated and edited images may store prompts, seeds, samplers, model references, XMP records, and provenance markers. Their presence is workflow evidence, not proof of AI origin.
Prompt / Negative prompt / Seed / Sampler / Steps / Checkpoint / LoRA / ComfyUI / AUTOMATIC1111 / C2PA
Images are re-encoded locally; PDFs are structurally rewritten for supported document properties. Each output is scanned again, and unsupported formats are rejected before processing.
Supported in V1
Images: up to 5 files. PDF: one file. 25 MB each.
Inspect hidden generation, provenance, and privacy data before you post, list, deliver, or send an image.
Check stored fields that may contribute to an AI label, then clean supported file-level metadata without claiming control over a platform’s classifiers.
Remove supported prompts, model and workflow references, GPS, and editor fields before publishing product, print-on-demand, or marketplace images.
Review prompt, model, author, copyright, software, and provenance fields before previews or final exports leave your team.
Remove supported location, device, identity, and workflow metadata before an image moves beyond a trusted circle.
Privacy should not depend on a promise. The browser network record and the output re-scan provide two separate checks.
0 file uploads
Analytics may send pageview requests, but file processing sends no file or metadata payload.
0 processing API
The cleaner calls no file-processing API.
0 file data sent
Only the disclosed page analytics contact external providers; file bytes stay local.
Clear answers about prompts, EXIF, image quality, provenance, AI labels, and the limits of file-level metadata cleaning.