Image Metadata Stripper

Introduction:

Image metadata is extra information stored beside the visible pixels. A phone photo may carry camera model, capture time, orientation, editing software, thumbnail data, comments, copyright fields, or GPS-like location text. Those details can be useful in a private archive and risky in a copy meant for public sharing.

Metadata cleanup is usually about separating the picture from the hidden record that travels with it. A clean copy should preserve the visible image while dropping common EXIF, XMP, IPTC, text, comment, and container markers that can reveal more than the viewer needs.

Removing metadata does not change what the photograph visibly shows. A street sign, badge, face, screen reflection, or distinctive background can still disclose private information. Metadata stripping is one privacy step before sharing, not proof that the image content itself is safe.

A good cleanup result also needs a verification step. Some formats expose clear metadata containers, while others depend on browser decoding support or only allow a limited scan. The safer reading is to check both the source markers and the clean copy markers before sending the image onward.

Technical Details:

Image metadata is stored in container structures around the pixel data. JPEG commonly uses application segments for EXIF, XMP, IPTC, JFIF, Adobe transform data, color profiles, and comments. PNG uses chunks, including text, EXIF, color, density, gamma, transparency, and other ancillary chunks. WebP uses chunks such as EXIF, XMP, ICC profile, and animation records. GIF can carry comments, application extensions, and frame-control data.

The practical distinction is between privacy-bearing metadata and technical display metadata. Privacy-bearing fields can include location hints, capture history, creator names, rights notes, editing software, document comments, and descriptive text. Technical metadata can include density, color profile, gamma, alpha, and animation structure. Technical markers are not automatically sensitive, but they can affect repeatability, display, and file-size comparisons.

Transformation Core

The cleanup path is a decode-and-re-encode process. The source image is read as bytes for marker inspection, decoded into pixels by the browser, drawn onto a canvas, and written as a fresh JPEG, PNG, or WebP file. Because the clean copy is made from pixels rather than copied container segments, common source metadata containers are not intentionally carried forward.

source image bytes
  -> inspect common JPEG, PNG, WebP, GIF, and generic metadata signatures
  -> decode one browser-readable image into pixels
  -> apply source orientation when selected
  -> draw pixels to canvas with JPEG background handling when needed
  -> encode clean JPEG, PNG, or WebP
  -> inspect clean output markers and build receipt rows

Canvas export has important boundaries. PNG output is lossless for decoded pixels, while JPEG and WebP use a quality value from the browser encoder. JPEG has no alpha channel, so transparent pixels are flattened onto the chosen background color. Animated inputs are decoded as one still image frame, and very large images can exceed safe browser memory limits.

Everyday Use & Decision Guide:

Start with the image you plan to share, not a resized screenshot of it. Drop, paste, or browse one JPEG, PNG, WebP, GIF, AVIF, BMP, or other image the current browser can decode. If several files are selected, only the first image is processed and the page shows how many extras were ignored.

For a first pass, choose JPEG when broad compatibility matters and the image has no important transparency. Use a quality value around 88 to 94 for typical sharing because it avoids the worst compression artifacts without preserving the original container. Choose PNG when transparency or lossless pixel output matters, and choose WebP when the recipient or publishing path supports it.

• Keep Apply source orientation on for phone photos, especially when the original depends on an Orientation tag to appear upright.
• Set Transparency background before JPEG export when the source has transparent edges or overlays.
• Open Metadata Findings when Sensitive metadata blocks, GPS hint, or Container needs review.
• Use Cleanup Receipt to confirm pixel dimensions, output format, size change, orientation handling, and browser-local processing.
• Use Size Cleanup Chart as a size and privacy-marker comparison, not as an image-quality score.

The main wrong assumption is that a smaller output file is always cleaner or better. Metadata removal can shrink a photo, but changing from PNG to JPEG or adjusting WebP quality can also change file size. Check Output privacy clean, Metadata Findings, and the visible preview before using the size number as a decision cue.

Use a stricter external parser when the stakes are high, such as legal discovery, newsroom source handling, compliance review, or a production privacy gate. The browser marker scan is useful for normal sharing review, but it is not a substitute for a full metadata audit across uncommon image containers.

Step-by-Step Guide:

Work from source selection to marker review so the clean image and the receipt describe the same run.

1. Drop, paste, or browse one image in Source image. The dropzone should move from Drop image here to Scanning and cleaning image..., then Clean image ready.
2. Choose Clean output format. Pick JPEG, PNG, or WebP based on the recipient and whether transparency must survive.
3. Adjust Quality for JPEG or WebP. PNG hides the quality control because the browser writes lossless PNG pixels.
4. For JPEG output, set Transparency background before trusting the preview. Transparent pixels are flattened onto that color.
5. Leave Apply source orientation on unless you intentionally want raw encoded orientation. If the clean copy looks rotated, switch this setting and compare again.
6. Read the summary badges. Output privacy clean means the clean copy scan found zero privacy markers; Review output markers means the output still needs inspection before sharing.
7. Open Metadata Findings and compare Source evidence with Clean output. A GPS hint in the source should become No GPS hint in the clean copy.
8. Open Cleanup Receipt for file size, dimensions, output format, orientation handling, and residual limitation notes. If the source is over 40 MB or cannot be decoded, choose a smaller browser-readable image.
9. Use Clean Image to inspect and download the finished copy, then keep JSON only when you need a structured record of the current settings, source inspection, output inspection, and warnings.

Interpreting Results:

The strongest privacy cue is the combination of Output privacy clean in the summary and zero privacy markers in Metadata Findings. If Sensitive metadata blocks changes from a positive source count to 0 privacy markers, the common marker scan agrees that the re-encoded copy dropped the scanned privacy containers.

GPS hint is a text-based clue, not a complete geolocation audit. It looks for GPS-like words in marker labels and sampled metadata text. A clean result lowers the risk that location tags remain in common metadata containers, but it does not hide location clues that appear inside the visible photograph.

Technical markers should not be read like privacy markers. A clean JPEG may still contain JFIF density data, a color-transform marker, or other technical output from the browser encoder. That can matter for display and repeat testing, but it is not the same as finding EXIF, XMP, IPTC, or comment content.

A successful browser cleanup does not prove that every metadata parser in the world would find nothing. Before a sensitive release, scan the downloaded copy with an independent metadata parser and visually review the image for faces, screens, place names, reflections, and other content that metadata removal cannot fix.

Worked Examples:

Phone JPEG With Location Tags

A phone photo is loaded as a JPEG, Clean output format stays on JPEG, Quality is set to 92, and Apply source orientation remains on. Metadata Findings may show Sensitive metadata blocks and GPS hint in the source. After cleanup, the expected review target is 0 privacy markers and No GPS hint for the clean output, with the preview still upright.

Transparent PNG Prepared For Sharing

A PNG logo with transparent edges is loaded and the output is changed to JPEG. The Transparency background color decides what replaces the alpha channel, so a white background may be correct for a document but poor for a dark site. Cleanup Receipt should report the chosen JPEG format and the pixel dimensions, while the preview shows whether the flattened background is acceptable.

Animated GIF Becomes A Still Copy

An animated GIF is loaded to remove comments or application extension metadata. The marker scan can report GIF comments, application extensions, or frame-control data, but the canvas export writes one decoded still frame. The warning about animated image input means the clean copy is useful for a still image handoff, not for preserving animation.

Large Or Unsupported Source

A 55 MB scan is rejected before cleanup because the browser-side file limit is 40 MB. A separate panorama under 40 MB may still fail if its decoded dimensions exceed 60,000,000 pixels. In either case, the corrective path is to downsize or pre-process the image, then reload it and confirm that Clean image ready appears before checking Metadata Findings.

FAQ:

Glossary:

EXIF

Image metadata often written by cameras and phones, including capture details, orientation, and sometimes GPS-like fields.

XMP

An extensible metadata format used for descriptive, rights, workflow, and editing information in many media files.

IPTC

Photo metadata fields often used for captions, creator information, rights, and editorial workflows.

GPS hint

A marker-scan clue that sampled metadata text contains GPS, latitude, longitude, geotag, or location wording.

Canvas re-encode

The browser process of drawing decoded pixels to a canvas and writing a fresh JPEG, PNG, or WebP copy.

Technical marker

Format, color, density, animation, or display metadata tracked separately from privacy-bearing fields.

References:

• About Exif 3.0, Camera and Imaging Products Association, December 2023.
• IPTC Photo Metadata Standard, IPTC.
• XMP Specifications, Adobe.
• Portable Network Graphics Specification, W3C.
• HTMLCanvasElement toBlob method, MDN Web Docs, last modified February 12, 2026.