Remove EXIF without uploading: verify GPS tags and the PNG result

Written by Feroz Sheikh · Founder, Doha, Qatar

Updated 2026-09-30. This is a working note for a live tool, not legal, tax, or professional advice.

Photo metadata can contain camera information, timestamps and location tags that are not visible in the picture. OpenToolSuite's EXIF and GPS Scrubber reads a selected image locally, draws its decoded pixels into a canvas and exports a fresh PNG. This guide documents a measured synthetic JPEG containing known tags. It also distinguishes removal of source metadata from hiding sensitive details in the visible scene, which remains a separate editing task.

The test image is a simple illustration rather than a real photograph. Its JPEG includes six top-level EXIF tags, including a synthetic camera label, a fixed date, a comment marker, orientation and a GPS directory with invented coordinates. Those values were added intentionally by the public fixture recipe. A known input matters: testing a photo with no metadata would not demonstrate that any tags were removed. The fixture contains no person's image or actual location information.

Choose the image already stored on your device, open the scrubber and select it through the file picker. Tap Scrub and download. The output uses a filename beginning with scrubbed and ends in .png. That changed format is deliberate. This workflow does not modify the original JPEG in place or simply remove its EXIF segment. It creates a new raster image from the browser-decoded picture, so you must keep the original separately if you need it later.

Canvas export is the mechanism behind this workflow. The browser decodes the selected local image, draws it into a canvas with the source dimensions and creates a PNG Blob from that canvas. The original EXIF data is not copied into that export by the tool. A browser can add information of its own to a generated file, so the accurate claim is about removal of the source metadata rather than an absolute guarantee that every output has no metadata of any kind.

In the recorded experiment, the real scrubber workspace exported the synthetic JPEG while the prepared browser context was offline. We then inspected the downloaded PNG with Pillow. The evidence table reports dimensions, source tag count, output tag count and whether the known GPS and comment marker remained. Those are checks of the specific output. The public JSON report retains input and output hashes and browser details, while the recipe lets you create the same controlled input yourself.

Open the downloaded PNG and review its visible content. Check orientation, colour, sharpness and transparent areas if your source includes them. A JPEG decoded and saved as PNG is a different file even if it looks similar. The output may be larger because PNG stores image data differently. This tool's primary aim is a clean local copy, not the smallest possible attachment. Do not assume a metadata scrubber also optimizes dimensions or meets a social platform's file limit.

A successful EXIF test does not hide clues in the pixels. A street sign, house number, school uniform, face, reflection or document visible in the frame can still reveal information after all source tags are removed. Inspect the whole scene before publishing. Crop or cover sensitive areas with an appropriate image editor and verify the edited export separately. For documents, simple visual edits should not be described as secure PDF redaction or metadata removal.

The filename can also reveal something you do not intend to share. A clean PNG named after a private address or account reference still carries that name when attached to a message. Rename the finished copy sensibly and choose its sharing destination deliberately. Removing EXIF cannot control what a messaging application, a cloud backup provider or a social network later adds, retains or associates with your account. Those services belong to a different part of the workflow.

To repeat the metadata check, use an image you know contains EXIF and GPS tags, or generate the public synthetic fixture. Inspect it before processing with a local metadata reader, then inspect the exact output file rather than a thumbnail copied by another app. Check both the tag directory and the known test marker. A side-by-side visual comparison alone cannot establish that hidden GPS fields are absent. The report includes these separate content and metadata observations.

The benchmark's offline result demonstrates that this prepared desktop workflow completed without an online image-conversion dependency. It does not claim a cold offline launch, physical iPhone testing or Android testing. The initial page and code were loaded before disconnection. It also does not audit every browser extension or future production configuration. If you require a broader privacy assurance, inspect the relevant build and test the exact browser and storage workflow you intend to use.

Supported formats depend on what the browser can decode. A file selected through image/* is not automatically supported; HEIC, unusual RAW formats or damaged pictures may fail. Convert only when you understand the effect on the image and metadata. Extremely large images may require substantial memory for the canvas. If the page cannot read the image or save a complete result, retain the source and use a trusted local image application rather than treating a failed download as cleaned.

Use the scrubbed copy only after verifying the result you will actually share. Confirm that the synthetic or real metadata fields you intended to remove are absent, review visible identifying details and rename the output if needed. The measured fixture shows a narrow, reproducible before-and-after case. It does not certify anonymity. Privacy comes from checking the saved file and the whole publishing workflow, including its pixels, filename, storage and destination.

Reference: MDN: canvas toBlob export

Reference: Pillow: Image EXIF inspection API

Reference: Manual checks and reproduction commands

Reference: Deterministic fixture generator

Reference: Local output verification recipe

Reference: Synthetic EXIF and GPS input

Reference: Measured scrubbed PNG

Open EXIF and GPS Scrubber

All guides