Frequently asked questions

Digital photos often carry invisible data embedded in the file: EXIF (camera settings, date taken), IPTC/XMP (creator, copyright), GPS coordinates, device make/model, and increasingly C2PA Content Credentials (AI provenance manifests). MetaPurge scans for all of these and can strip them before you share an image.

The small "CR" icon on LinkedIn images indicates Content Credentials — hidden C2PA metadata LinkedIn detected in the file. It does not necessarily mean the image was AI-generated; many editing apps embed C2PA by default. Standard cleaning removes the manifest that triggers the badge.

Three steps: (1) ExifTool deletes every metadata tag it can reach with -all=. (2) Sharp re-encodes the pixel data into a fresh file container — metadata is not carried forward unless explicitly requested. (3) A byte-level scanner verifies no C2PA/JUMBF markers survived. Your file is processed in a temp directory and deleted immediately after the response — nothing is stored.

No. Files are written to a temporary directory for the duration of a single request, then deleted in a finally block. We never persist uploaded images or their metadata content to disk or database.

Free — 5 cleans per day (tracked by IP, no signup), single file up to 40MB, metadata summary report, unlimited scan-only (viewer) mode. Pro ($5.99/mo) — unlimited cleans, batch upload with ZIP download, up to 1GB per file, full metadata viewer (every field), and Deep Purge. See pricing.

Deep Purge is a Pro-only optional pass applied after the standard metadata strip. It runs a subtle pixel-level step: a brief 98% resize round-trip followed by a fresh re-encode at slightly reduced quality. The goal is to disrupt some pixel-level fingerprinting that a metadata-only strip does not touch. It is not a guarantee against invisible watermarks (e.g. SynthID) or ML-based AI image classifiers that analyze pixel content rather than embedded metadata. Use it when you want an extra layer beyond standard cleaning — not as a promise of undetectability.

After scanning, Pro users see every detected field listed individually (GPS, device info, C2PA, etc.). Free users receive a condensed summary (found/not found for key categories). Both tiers run the same inspect pass — Pro just gets the full report rendered in the UI.

Scan only (viewer mode) inspects metadata without modifying the file — useful for checking what is embedded before deciding to clean. It does not count against your daily clean limit. Scan & clean strips metadata and returns a downloadable clean file (counts toward Free tier daily cap).

HEIC/HEIF input is re-encoded as JPEG on output for broad compatibility. The cleaned file will have a .jpg extension even if the original was HEIC. Image quality remains high (95% JPEG quality on the standard pass).

Metadata and C2PA removal defeats metadata-based labeling (e.g. LinkedIn's CR badge). It does not remove invisible pixel watermarks baked into image data (e.g. Google SynthID-style systems) or defeat ML classifiers that analyze pixels rather than embedded tags. If a platform still flags an image after cleaning, the cause is likely outside what metadata stripping can address.

Subscriptions are managed through Stripe. Cancel from your Stripe customer portal or contact us via the contact form. After cancellation, your account reverts to Free tier limits at the end of the billing period.