EXIF Forensics Tool Private · No Upload
100% privat EU DSGVO Nichts verlässt deinen Browser.

Fotos. Privat. — Free EXIF Viewer and Metadata Remover. Strip GPS, camera info, dates and software traces from photos, videos, audio, PDF and Microsoft Office files. 100% in your browser, no upload, fully anonymous.

Entferne GPS, Kamera-Daten und Spuren aus deinen Fotos, Videos, PDFs und Office-Dateien. Direkt im Browser.

Kein Upload Keine Cookies Kein Account 100 % kostenlos

In jedem Foto.

Mehr als das Bild.

Kameras und Smartphones speichern bei jedem Foto unsichtbar mit, wo es entstanden ist, mit welchem Gerät, zu welcher Sekunde — manchmal sogar die Seriennummer.

📍
GPS-Position52.5163° N · 13.3777° E
📷
KameraiPhone 15 Pro · 48mm
📅
Aufnahme2026-04-15 · 14:22:33
🔢
SeriennummerF2LX9K7P-22

So einfach.

Drei Klicks zum sauberen Foto.

1

Hineinziehen.

Foto, Video, PDF oder Office-Datei. Ein einziger Drop genügt — oder klicke und wähl mehrere aus.

2

Profil wählen.

Anonym, soziale Medien, Kamera behalten — vier vorgefertigte Karten. Du musst nichts wissen.

3

Saubere Datei.

Klick auf den Knopf. Die saubere Datei landet sofort in deinem Download-Ordner.

Jetzt ausprobieren.

Hier deine Datei.

Drop files here, click to select, or paste with Ctrl-V

Photos · Videos · Audio · PDF · Office (DOCX/XLSX/PPTX) · Multiple files · Drag from any webpage · Paste from clipboard

Datei wählen

No photo handy?

Sample image credit & license

The sample image is based on Cat03.jpg by Alvesgaspar via Wikimedia Commons, licensed under CC-BY-SA 3.0. The cat was not actually photographed at the Brandenburg Gate — the GPS coordinates (52.5163° N, 13.3777° E) were synthetically added on top of the file purely so this demo shows an unmistakable, world-famous location. That makes the privacy point immediately obvious: a real photo could just as easily reveal your home, your office or a child's school with the same precision.

Sicher.

Niemals auf einem Server.

Andere Tools laden deine Datei hoch und versprechen, sie zu löschen. Hier passiert nichts dergleichen. Drücke F12, öffne den Network-Tab — du wirst keinen Upload sehen. Trenne deine Internetverbindung — das Tool funktioniert weiter.

🔒
Lokal verarbeitetBrowser File API liest die Bytes — sie verlassen dein Gerät nie.
🛰️
Funktioniert offlineLade die Seite einmal, schalte WLAN aus — alles arbeitet weiter.
📖
Quellcode lesbarView Source. Keine Minifizierung. Suche nach „fetch" — du findest nichts.

Alles dabei.

Fotos, Videos, Dokumente.

📷 Fotos JPG · PNG · WebP · HEIC · TIFF · GIF · BMP · PSD · ICO
🎬 Videos MP4 · MOV · MKV · WebM · AVI
🎵 Audio MP3 · FLAC · OGG · WAV
📄 Dokumente PDF · DOCX · XLSX · PPTX

Why we built this — and why it's truly anonymous

We built this tool because most online metadata removers do exactly what you should avoid: they ask you to upload your private files — photos, videos, audio, PDFs, Word and Excel documents — to a server you don't control. That's the opposite of privacy. Even when those services promise to delete your file afterwards, you have no way to verify it, and your file already passed through their network, log files, error monitors, and possibly backups.

This tool is different. Your file never leaves your computer. Whether you drop a holiday photo, a recorded interview, a confidential PDF or an Office document, the browser opens it locally, parses the metadata locally, removes what you choose locally, and lets you save the cleaned file locally. There is no upload step at any point.

You can verify this yourself: press F12, open the Network tab, and drop a 50 MB file (a video or a large DOCX works just as well as a photo) into the tool. You will see no upload request. The tool only loads its own JavaScript and CSS files once — after that, it works fully offline.

  • No upload — files are read with the browser File API, never sent anywhere
  • Your photo is never tracked — your image data never enters any analytics, ever
  • No account, no email, no cookies for the tool — your language preference is stored locally, that's all
  • No GPS lookup — we never query a map service with your coordinates; we only display them as text
  • Auditable — view-source on this page shows exactly what runs
  • Works offline — once the page is loaded, you can disconnect from the internet

Who this tool is for

This tool serves anyone who needs to inspect, clean or verify a file without sending it to a third-party server. Six groups use it most:

Privacy-conscious individuals

Strip GPS coordinates from holiday photos before posting to Instagram, eBay or a dating profile. Clean a CV photo, a marketplace listing, or a screenshot before sending. Find out what an existing photo from an ex-partner, stalker or unknown sender already reveals about you. Remove device-serial numbers from photos before public posts. Free EXIF cleaner, GPS remover, metadata scrubber — all in your browser, no upload, no tracking.

OSINT researchers & fact-checkers

Open-source intelligence community, Bellingcat-style investigators, fact-checking desks at BBC Verify, dpa, AFP, AP, Reuters and Snopes. Run Error-Level-Analysis, JPEG-Ghost, CFA / Bayer-pattern checks, Copy-Move detection, DQT-fingerprint matching, RGB-histogram gap analysis and pixel-perceptual-hash thumbnail comparison. Generate signed C2PA Content Credentials for chain-of-custody. Export a printable forensic report. The full FotoForensics + Forensically toolkit, but without the upload — important when investigating leaked, sensitive or jurisdictionally-tricky material.

Journalists & newsrooms

Verify user-generated content (UGC), citizen-journalism submissions and crowd-sourced photos before publication. GDPR-compliant: nothing leaves the editorial laptop, so source-protection rules and works-council agreements stay intact. Useful for local newsrooms without an in-house forensics desk, and for investigative reporters working on cross-border stories where uploading to US servers would be a problem.

Lawyers, investigators & loss adjusters

Pre-screen evidence photos in family-law cases, asset disputes and insurance claims. Document a client's photo in a confidential intake interview without uploading it anywhere — attorney-client privilege stays watertight. Generate a PDF forensic report with SHA-256 hash and timestamp as an internal case-file note. Not court-admissible by itself (use accredited tools like Amped Authenticate for that), but excellent as a first triage step before commissioning expert analysis.

Education & training

Media-studies and digital-forensics courses at universities. Police OSINT training, journalism schools (Henri-Nannen, DJS, Columbia J-School), IT-security workshops. Free of charge, multilingual (19 languages), runs offline, beginner-friendly explanations next to every forensic technique. Students see real ELA heatmaps, Copy-Move clusters and DQT fingerprints on their own files instead of textbook examples.

Activists, NGOs & human-rights orgs

Human-rights documentation in conflict zones (Amnesty International, Human Rights Watch, Witness.org, Bellingcat-style war-crimes investigators). Whistleblower-protection workflows where uploading evidence to a US-hosted analyser would be a hard veto. Verify atrocity-footage from Telegram, Mastodon, X. Source-protection desks at platforms like SecureDrop and GlobaLeaks. Field-laptops in low-bandwidth regions: works fully offline once loaded. The privacy guarantee here is not marketing — it is operationally critical.

Common thread: everyone who needs metadata insight without giving up control of the file. If your workflow says "do not upload to third-party servers" — for legal, ethical, contractual or personal reasons — this tool fits.

What is metadata?

Metadata is hidden information that your devices and software automatically embed inside every file you create — photos, videos, audio recordings, PDFs and Microsoft Office documents. The best-known kind is EXIF (Exchangeable Image File Format) inside photos, but every common file type carries its own variant. Typical content includes:

  • GPS coordinates — the exact location where the photo or video was taken, often accurate to within a few meters; smartphones also write GPS into MP4/MOV recordings
  • Date and time — when the file was created or last edited, often with timezone
  • Camera / device model and serial number — the hardware that produced the file
  • Lens, exposure, ISO, aperture — full technical settings for photos
  • Editing software — the program used to create or modify the file, often with exact version
  • Author / Owner / Last modified by — names and copyright info embedded by the camera, recorder or Office
  • Direction, altitude, speed — on smartphones and recording devices with motion sensors
  • Audio tags (ID3, Vorbis comments) — artist, album, encoder, embedded album art, sometimes lyrics or private DJ-software frames
  • Document properties — Word/Excel/PowerPoint store author, company, manager, total editing time, revision counter and template path inside every document
  • PDF info — Title, Author, Subject, Keywords, Producer (the software that wrote the PDF), plus an XMP metadata stream that Adobe Reader prefers over the basic info dict

This information travels with the file wherever it goes — by email, messenger, forum, file-sharing service or website upload — unless someone removes it. Most people never realize their files carry it.

Why does this matter?

Most people have no idea what their files quietly carry. Metadata is invisible in normal viewers — it only surfaces when a stalker, an investigator, an abusive ex, a curious recruiter, an advertising broker, an OSINT-trained doxer or a hostile state actor opens the file with the right tool. By then it is too late. Below is a non-exhaustive list of real, documented consequences that have already happened to real people:

  • Home address exposure — a single GPS-tagged photo of your living room, your driveway, or even just a window in the background reveals exactly where you live, often accurate to a few meters. Real-estate listings, second-hand-marketplace photos, dating-app pictures and "look at my new place!" posts have all been used to locate people
  • Stalking, harassment and domestic abuse — fleeing victims of domestic violence have been tracked down by ex-partners through GPS in children's photos posted to family groups or public profiles. Shelters explicitly warn residents to strip metadata before sharing any image
  • Anonymous sources deanonymized — bloggers, whistleblowers, journalists and activists have been exposed by metadata in their own files: GPS in protest photos, author names in leaked Word documents, camera serial numbers correlating images across years, "last modified by" fields on contracts
  • Pattern-of-life surveillance — timestamps and GPS together produce a calendar: when you are home, at work, on vacation, at the gym, at the doctor, at someone's apartment overnight. Five photos are usually enough to map a life
  • Device fingerprinting — your camera's serial number, lens serial number and software fingerprint travel inside every photo. Researchers can correlate "anonymous" images posted years apart and across different platforms back to the same device — and from there, to you
  • Office-document leakage — Word, Excel and PowerPoint files secretly carry author names, "last modified by" of every co-editor, total editing time, revision counters, hidden tracked changes and the company-template path. Job applicants have been screened out by recruiters who saw the previous applicant's name still in the document properties; whistleblowers have been identified by Office’s built-in author field
  • PDF Producer fingerprinting — every PDF declares which software wrote it (Producer, Creator) and often a full XMP metadata stream. Anonymous reports, leaked legal filings and "untraceable" tip-offs have been sourced back to specific organisations through these fields
  • Audio file deanonymization — MP3 ID3 tags can contain the recording device, the editor (Audacity, Adobe Audition with login name), embedded album art with its own EXIF, and time-of-encoding. Voice memos and amateur podcasts have leaked identities this way
  • Video file location and equipment leak — modern smartphones write GPS into MP4/MOV recordings of every video you take. MKV/WebM exports from professional editors carry the editor name, the encoding host’s name and the writing application. A "random" street video can pinpoint where you stood, when, and which device you used
  • Forensic correlation across platforms — the same camera serial number, software signature or document fingerprint appearing on a dating-app photo, a Reddit post, a leaked PDF and a "stranger’s" Wikipedia upload effectively merges those identities. This is how the FBI links suspects, how journalists confirm sources, and how harassers find their targets

Two famous cases worth knowing: John McAfee, 2012 — while a fugitive in Central America, a Vice journalist photographed him with an iPhone and uploaded the picture with intact GPS metadata. Within hours, anyone reading the article knew his exact hideout in Guatemala. Higinio O. Ochoa III, 2012 — an Anonymous-affiliated hacker posted an anonymous selfie of his girlfriend taunting law enforcement; the FBI extracted GPS from the iPhone EXIF and used it to locate and arrest him. In both cases the public photo carried the smoking gun the suspect did not realise he was sharing.

Even if you are neither a fugitive nor an activist, the same mechanics apply on a smaller scale every day: a sold-house photo on a marketplace, a child’s school-event picture in a chat group, a CV opened by a stranger, a video thumbnail uploaded to a forum. The data is already in the file. The only question is whether you remove it before sharing — or whether the next person to open the file is the one who finds it first.

How to use this tool

  1. Drop your filesDrag one or more files into the upload area, or click to browse. Photos (JPG/PNG/WebP/HEIC/TIFF/GIF/BMP), videos (MP4/MOV/MKV/WebM/AVI), audio (MP3/FLAC/OGG/WAV), PDF and Microsoft Office (DOCX/XLSX/PPTX) are all supported.
  2. Review what's insideFor each file, the tool lists every metadata tag found, grouped by category. GPS coordinates are highlighted at the top.
  3. Choose what to removeClick 'Remove all' for a fully anonymous photo, or use category buttons to keep some metadata (e.g. keep camera info, strip GPS).
  4. Save the cleaned photoThe download contains the same image at the same quality, just without metadata. Use the ZIP download for batches.

Features

100% Private No upload — your photo is processed entirely in your browser using the File API
All major formats JPEG, PNG, WebP, TIFF, HEIC, GIF, BMP photos · MP4, MOV, MKV, WebM, AVI video · MP3, FLAC, OGG, WAV audio · PDF · Microsoft Office DOCX, XLSX, PPTX — covering virtually every file type that carries metadata
GPS detection Latitude and longitude shown in degrees, with optional one-click conversion to decimal
Batch processing Drop multiple files at once, review individually, download all as a single ZIP
Selective stripping Remove only GPS, only date, only software — or everything in one click
Lossless Removing metadata does not re-encode your image — pixels and quality are unchanged
Smaller files Cleaned photos are typically a few kilobytes smaller — useful for email and uploads
Truly free No ads, no third-party tracking, no account, no email, no upload limits, no premium tier
19 languages English, Deutsch, Español, Français, Italiano, Nederlands, Polski, Português, Русский, Українська, Türkçe, العربية, हिन्दी, বাংলা, 中文, 日本語, 한국어, Bahasa Indonesia, Tiếng Việt
JSON export Download all parsed metadata as a structured JSON file — for forensic reports, scripted analysis or archival before stripping. Generated in your browser, never uploaded.
SHA-256 hash Each file is hashed locally via WebCrypto so you can verify the exact bytes you analysed — useful for chain-of-custody documentation and de-duplication.
ICC profile details Profile name, color space, device class and manufacturer extracted instead of just '[present]' — see whether your photo carries sRGB, Display P3, Adobe RGB or a custom calibration.
Camera RAW support Reads CR2, NEF, ARW, DNG, ORF, PEF, RW2 (TIFF-based RAWs), plus Canon CR3 and Fujifilm RAF — view-only, exact same EXIF parser as JPEG.
One-click privacy profiles Four named presets — Anonymous, Social media safe, Keep camera info, Just remove GPS — match common situations so you don't need to know which checkboxes to tick.
Per-tag EXIF editor Change Make, Model, Software, Artist, Copyright, Description, DateTime and Orientation directly in JPEG without re-encoding the image.
GPS spoofing Replace real coordinates with a famous landmark or any custom lat/lon. Useful when platforms flag photos with no GPS as suspicious.
MakerNote decoder Optional Deep Inspection unwraps the proprietary MakerNote tag for Canon, Nikon and Sony — surfaces lens model, shutter count, firmware, internal serial.
ExifTool-compatible export Toggle the export format to ExifTool's flat -j JSON schema and drop our exports into existing forensic / OSINT pipelines without writing a translator.
CSV export Export all metadata as a flat CSV that opens directly in Excel, LibreOffice or Google Sheets — perfect for spreadsheet-based audits.
Tag-diff after stripping After every clean operation, see exactly which tags were removed and which were kept — concrete confirmation, not just a "size saved" number.
Filename sanitiser Auto-removes date/time/coordinate patterns from camera-generated filenames (IMG_20240315_142233.jpg → image-clean.jpg) before the download.
Hex viewer for every format Inspect raw bytes of EXIF / PNG chunks / MP4 boxes / MKV elements / PDF trailer — the same forensic view you'd get from a hex editor, in the browser.
Privacy Risk Score (0-100) Traffic-light score that quantifies how identifiable your file is, with a per-file breakdown showing exactly which metadata items pushed it up.
Plain-language tag explanations Click the "?" next to any tag to learn what it means, why it's privacy-sensitive, and what someone could do with it. Bundled — no server contact.
Date timeline All dated tags plotted on a single chronological axis. A genuine photo's dates cluster within seconds — a forged one shows wide gaps.
Embedded thumbnail audit Detects the IFD1 preview most strippers forget — and warns when it carries its own GPS ("ghost GPS" leak).
Tampering hints Cross-references date fields, software signatures and MakerNote vendor against camera Make. Surfaces forgery-style inconsistencies forensic analysts look for.
ExifTool CLI snippet Generates the equivalent ExifTool command for every operation — useful for scripting, batching, and verifying our output against the reference implementation.
Compression fingerprint Matches the JPEG quantization tables against twelve known Photoshop "Save for Web" profiles plus the libjpeg standard. Identifies the software and likely pipeline (capture → editor → messenger). Read-only, runs locally.
Structure inspector Walks every JPEG marker, PNG chunk and HEIC/MP4 box as a clickable colour-coded list. Detects polyglot files (data hidden after EOI/IEND), exotic OEM markers and steganographic containers.
JPEG-Ghost heatmap Re-saves the file at eight quality levels and finds the best-matching Q for every 16×16 region. Inserted regions from a different-quality source stand out as a coloured cluster — Hany Farid's classic 2009 forensic technique.
Noise + Bayer pattern map Two pixel-level checks. Noise variance flags suspiciously smooth blocks (denoising, retouching, clones). The Bayer-sensor footprint score (0–100) is a defensible alternative to black-box AI detectors — synthetic images score near 0, real photos 50+.
Copy-move detection Hashes every 16×16 block and finds groups of identical-looking blocks shifted by the same amount — the unmistakable signature of clone-stamp retouching. Runs in a Web Worker so the page stays responsive.
Thumbnail-Mismatch detector Compares the JPEG's embedded IFD1 thumbnail with a downsampled version of the main image via 64-bit perceptual hash. After Photoshop edits the thumbnail is often not updated — different content = near-bulletproof manipulation proof. Hany Farid 2009 classic.
RGB histogram + spike detection Plots the per-channel R/G/B brightness distribution and counts empty bins inside the active range. Levels, curves and tone-mapping leave characteristic gaps — over 20 empty bins is almost always a sign of Photoshop/Lightroom adjustment.
Pixel magnifier (8× zoom) Hover or long-press on any forensic heatmap and a zoom-bubble follows your cursor with 8× magnified pixels, RGB values and hex colour code of the centre pixel. Essential for reading ELA, Noise and CFA maps at the pixel level.
Forensic report PDF export Click one button — a print-ready document opens with file hash, EXIF highlights, compression fingerprint and every forensic finding (ELA, Ghost, Noise, CFA, Copy-Move, Thumbnail, Histogram). Browser's "Save as PDF" produces a clean evidence document. No PDF library, no upload — everything client-side.
Side-by-side compare slider Drop two images, drag the slider in the middle and reveal differences pixel by pixel. Classic before/after forensic tool for spotting Photoshop edits, comparing same-scene shots and tracking compositing seams. Pixel magnifier works on both halves.
C2PA Self-Sign Credentials Stamp any JPEG with a digital "seal of authorship" — like signing a letter with your own pen. Browser generates a fresh cryptographic key, signs the photo, throws the key away. The signed JPEG can be verified at contentcredentials.org. The only browser-only C2PA tool that doesn't upload your file. Self-signed (no commercial CA), but mathematically valid.

How does this compare to other EXIF tools?

Most online EXIF tools fall into two camps: server-based services that ask you to upload your private photos, and browser-based tools with limited features or aggressive monetisation. Here is how this tool stacks up:

Reading the table: green ✓ marks an answer that is good for your privacy; red ✗ marks one that is bad for you. The same word (e.g. "rarely") can therefore appear in both colors depending on the question — for example "Works offline → rarely" is bad (red), but "Account required → rarely" is good (green).

This tool Server-based tools Other browser tools
File leaves device ✗ no ✓ yes ✗ no
Source auditable ✓ yes ✗ no usually
Multilingual ✓ 19 languagesEN only occasionally
Works offline ✓ yes (PWA) ✗ no ✗ rarely
Formats supported ✓ photos + video + audio + Office + PDFmostusually JPG only
Audio metadata (MP3/FLAC/OGG/WAV)✓ yespaid only✗ no
Office files (DOCX/XLSX/PPTX) ✓ yespaid only✗ no
Batch processing ✓ ZIP downloadoften paywalledoften missing
Ads / Tracking ✗ none ✓ heavy often present
Account required ✗ none often✓ rarely
GPS map call (leak risk) ✗ never often often
JSON / structured export ✓ yes paid only ✗ rarely
SHA-256 file hash ✓ yes ✗ no ✗ rarely
Camera RAW (read) ✓ yes ✓ yes ✗ rarely
ICC profile details ✓ yes usually ✗ no
Privacy profiles (presets)✓ yes ✗ no ✗ no
Per-tag EXIF editor ✓ yes (JPEG) paid only ✗ rarely
GPS spoofing ✓ yes ✗ no ✗ no
MakerNote decoding (Canon/Nikon/Sony)✓ yespaid only✗ no
ExifTool-compatible export✓ yes ✗ no ✗ rarely
CSV export ✓ yes paid only ✗ no
Tag-diff after stripping ✓ yes ✗ no ✗ no
Filename sanitiser ✓ yes ✗ no ✗ no
Privacy Risk Score (0–100)✓ yes ✗ no ✗ no
Plain-language tag explanations✓ yes ✗ no ✗ no
Date timeline view ✓ yes ✗ no ✗ no
Embedded thumbnail audit✓ yes ✗ no ✗ rarely
Tampering / inconsistency detection✓ yes✗ no ✗ no
ExifTool CLI snippet generator✓ yes ✗ no ✗ no
TIFF strip (selective + full) ✓ yes ✗ no ✗ no
BMP strip (V5 ICC-color profile)✓ yes ✗ no ✗ no
Error Level Analysis (ELA) ✓ yes server-only✗ no
JPEG quality estimator ✓ yes server-only✗ no
DQT compression fingerprint + pipeline✓ yesserver-only✗ no
File structure inspector (markers/chunks/boxes)✓ yes✗ no✗ no
JPEG-Ghost re-save heatmap ✓ yes server-only✗ no
Noise variance + CFA / Bayer map✓ yesserver-only✗ no
Copy-move clone detection ✓ yes server-only✗ no
Thumbnail-Mismatch detection✓ yes✗ no✗ no
RGB histogram + spike detection✓ yes✗ no✗ no
Pixel magnifier (8× zoom + RGB)✓ yes✗ no✗ rarely
Forensic report PDF export✓ yespaid only✗ no
Side-by-side compare with slider✓ yes✗ rarely✗ no
C2PA Self-Sign provenance (browser-only)✓ yes✗ no✗ no
Cost free freemiumfree

Auditable means: open the .js files directly in your browser and read what they do. No minification, no obfuscation, no remote code execution.

Methodology: the "Other browser tools" column is a best-effort summary of the publicly visible features of exif.tools, onlineexifviewer.com, verexif.com, imageonline.co/exif-tool, removemyexif.com as of 2026-05. "Server-based tools" covers jimpl.com, metadata2go.com, pic2map.com, fotoforensics.com. We mark a feature "rarely" rather than "no" when at least one tool in the column has it. If you spot an inaccuracy, please email so we can correct it — we want this table to be defensible, not boastful.

FAQ

Is my photo really not uploaded?

Correct. Everything runs in your browser. You can verify it by opening the Network tab in your browser's developer tools (F12) and dropping a large photo — you will see no outgoing request that contains your file. After the page loads its own JavaScript and CSS once, the tool can run with the internet disconnected.

What metadata exactly is removed?

For JPEG: all APP1 (EXIF, XMP), APP2 (ICC, MPF), APP13 (Photoshop, IPTC) and APP14 (Adobe) segments. For PNG: tEXt, iTXt, zTXt, eXIf and Adobe XMP chunks. For WebP: the EXIF, XMP and ICCP chunks plus the VP8X presence flags. For GIF: Comment and Plain-Text extensions, plus XMP application extensions (NETSCAPE animation loop is preserved). For MP4/MOV: udta/meta/ilst boxes are neutralized to "free" atoms, and mvhd/tkhd timestamps are zeroed. For MKV/WebM: Tags, Attachments, Title, DateUTC, MuxingApp and WritingApp elements are replaced by Void elements of identical byte length. For AVI/WAV: LIST INFO and bext/iXML/IDIT chunks are renamed to JUNK. For audio (MP3, FLAC, Ogg): full ID3v1/v2/APEv2, Vorbis-comment + PICTURE, and CRC-correct page rewriting respectively. For PDF: DocInfo dict + XMP metadata stream override via incremental update. For DOCX/XLSX/PPTX: docProps/core.xml + app.xml + custom.xml replaced by anonymized stubs. The actual pixel data, audio samples, video frames and document body are never touched.

Does removing EXIF reduce image quality?

No. EXIF is a separate metadata block inside the file. Removing it does not re-encode the image — the pixels and the JPEG/PNG/WebP compression are exactly as they were. The file becomes a few kilobytes smaller because the metadata is gone.

Can I keep some metadata and remove the rest?

Yes. Each file card has buttons to remove only specific categories — e.g. "Remove only GPS" keeps everything except location, "Remove only Camera info" keeps everything except make/model/serial. The default "Remove all" wipes the entire metadata block at once.

Do social networks already remove EXIF?

Some do, some don't, and the rules change without notice. Facebook, Instagram and Twitter strip most EXIF on public posts but keep some internally for moderation and analytics. WhatsApp, Telegram (when "send as file"), Discord, email attachments, blogs, forums, eBay listings, classifieds, cloud drives — almost all of these keep EXIF intact. The safe rule is: never assume any service removes EXIF for you.

Why don't you show the GPS location on a map?

Because the moment we automatically loaded a map tile — even a tiny preview — your browser would send your photo's GPS coordinates to a third-party server (Google Maps, OpenStreetMap, Mapbox) without you doing anything. That third party would silently learn that someone with this IP just opened a photo carrying these coordinates — exactly the kind of side-channel leak this tool exists to prevent. So the tool itself never contacts any map service: latitude and longitude are shown as plain text only. We do, however, offer a clickable «Open in OpenStreetMap» link next to the coordinates — clicking it opens OSM in a new tab and reveals the coordinates to OSM. That request only happens because YOU chose to click; it is your decision, not the tool's, and you can ignore the link or copy the coordinates into any other map service instead.

What about HEIC photos from iPhones?

EXIF can be read from HEIC files in this tool. Stripping HEIC requires re-encoding the image, which most browsers cannot do natively. For full stripping of HEIC, the safest path is to first convert to JPEG using your operating system or photo app, then strip with this tool.

Is the tool open source?

The code is delivered to your browser as plain JavaScript files and is fully readable. Right-click → View Source, or open the .js files directly. There is no minification or obfuscation that hides what the tool does.

What's the catch? Why is this free?

There is no catch. The tool is provided as a free service by Midgard MUD — a free Norse-mythology online role-playing game. Some visitors discover Midgard through this tool and try the game, but you are not asked to do so and the tool works fully without any account or interaction with the MUD.

Can I use this on my phone?

Yes. The tool works in any modern mobile browser — Safari on iOS, Chrome and Firefox on Android. On iOS you may have to choose "Files" instead of "Photos" in the picker to access HEIC originals.

Why does my photo still show some date/dimension info?

The image dimensions (width × height) and the file size are inherent to the image itself, not metadata — they cannot be hidden without re-encoding. The "Date Modified" timestamp on the downloaded file is set by your operating system at the moment of saving and is unrelated to EXIF. To control that, set the file's modification date manually in your file manager after download.

Are there any limits on file size or number of photos?

There is no hard limit. Practical limits are set only by your device's memory: a typical phone or laptop comfortably handles 50+ photos at 20 MB each. For very large batches (hundreds of files), processing in two or three runs is more responsive than one giant batch.

Does the tool also remove watermarks from photos?

It depends on what kind of watermark you mean. Visual watermarks — the semi-transparent logo a photographer overlays on the image itself — are part of the pixel data, and removing them would require re-painting those pixels (which we don't do; the tool never re-encodes the image content). Invisible watermarks — steganographic identifiers embedded in the pixels by Adobe Stock, Getty, etc. — are likewise pixel-domain and untouched by metadata stripping. What we DO remove is everything that lives in the metadata block: copyright fields, IPTC author tags, XMP rights statements, "edited in Photoshop" producer strings, embedded Adobe credentials. So if a stock-photo site relies on EXIF-level rights management to identify the licensee, that information is gone after our strip — but the visible logo on the photo and any pixel-level steganography both remain.

I am a complete beginner — what is EXIF anyway, in one paragraph?

When you take a photo with a smartphone or camera, the device automatically writes a "data label" inside the photo file alongside the actual image. That label includes things like exactly where you were standing (GPS coordinates, often accurate to a few meters), the time and date, the device you used, even the device's serial number. The data is invisible when you view the photo, but anyone who receives the file can read it with a free tool — including the recipient of an email, a buyer on eBay, a forum reader, a stalker. EXIF stands for "Exchangeable Image File Format" and is the technical name for that label. Removing it before sharing a photo is a basic privacy habit, similar to closing the curtains before getting changed.

I just want to share a photo with friends — do I really need this?

Honest answer: probably yes, especially for photos taken at home. Common scenario: you take a photo of your new puppy in the living room, share it on a Telegram channel or post in a Facebook group. The photo carries GPS coordinates a few meters from your couch. Anyone in that channel/group — even people who don't know you personally — can extract those coordinates with a free tool, paste them into Google Maps, and see your front door on Street View. For photos of your kids at home this is a particularly real concern. WhatsApp strips EXIF when you send a picture as an image, but Discord, Telegram (file mode), email, blog posts and most other places keep everything intact. A 10-second strip before sharing closes that gap.

How do I know nothing is uploaded — without trusting your word for it?

Three independent ways to confirm. (1) Trust through verification: press F12 in your browser → "Network" tab → drop a large photo into the tool → watch the network log. If we uploaded anywhere, you'd see an outgoing request containing the file. You won't — the only requests visible are the tool's own JavaScript files, loaded once when you opened the page. (2) Disconnect and still works: load the page once with the internet on, then turn Wi-Fi off or enable airplane mode. The tool keeps working perfectly because everything runs locally. A server-based tool would fail immediately. (3) View source: right-click → "View page source", then click on "exif-tool.js" — the code is plain JavaScript, no minification or obfuscation. Search for keywords like "fetch" or "XMLHttpRequest" — you'll only find them in places that load translation files from your own browser, never to send image data anywhere. If you're still uncomfortable, you don't have to use the tool — that's a perfectly valid response to "trust me, it's safe" claims on the internet, and we respect it.

Does it work for audio files too?

Yes, and across four different audio formats with the same one-click flow you get for photos. MP3 files carry three independent metadata layers — an ID3v2 tag at the start of the file (Title, Artist, Album, lyrics, embedded album art, encoder strings, sometimes private GEOB frames placed there by recording or DJ software), an ID3v1 tag in the last 128 bytes, and occasionally an APEv2 footer used by Foobar2000 and similar players. The tool removes all three at once and leaves the actual MPEG audio frames bit-identical, so the audio quality is unchanged and the file is a few kilobytes (or several megabytes if cover art was embedded) smaller. FLAC files have their Vorbis-comment block (artist, title, replay-gain, copyright) and any PICTURE block stripped, but the STREAMINFO block stays intact so duration and sample rate remain visible to players. Ogg Vorbis and Ogg Opus are the most delicate: the comment payload is wrapped in Ogg pages with strict CRC32 checksums, so the tool replaces the comment packet with a minimal empty one, re-segments it, and recomputes every affected page's CRC32 (polynomial 0x04C11DB7) — strict players like Rockbox and libogg-based tooling continue to validate the file. WAV files get their LIST INFO entries (Artist, Comments), Broadcast-Wave (bext) chunks, iXML, embedded XMP and any inline ID3 sub-chunks turned into RIFF JUNK, which keeps the chunk offsets valid for editors. No audio is ever re-encoded.

Can it strip Word, Excel and PowerPoint metadata?

Yes — DOCX, XLSX and PPTX are all supported, and this is actually one of the most under-appreciated privacy use-cases. Microsoft Office stores a surprising amount of personal data inside every document: the author name and "last-modified-by" name, your company name, your manager's name (if set in Office options), the application version (revealing the exact Office build), the total number of minutes you actually spent editing the file, the revision counter, the template path, plus any custom user-defined properties. All of that lives in two small XML files (docProps/core.xml and docProps/app.xml) inside the ZIP container that an Office document really is. The tool unpacks the document in your browser, replaces those two XML files with anonymized stubs that keep the correct XML namespaces (so Office opens the file without a recovery dialog), wipes any docProps/custom.xml entries, and re-packs the ZIP. Crucially, the actual document content — your text, paragraphs, formulas, slides, embedded images, charts — is copied through byte-for-byte, so opening the cleaned file in Word, Excel or PowerPoint shows exactly the same content with no warnings. One thing worth knowing: tracked changes and comments inside the document body contain their own author attribution; those are content-bearing and not removed by the standard strip, so for a fully anonymous document you should accept all tracked changes and delete comments inside Office before downloading.

What is the SHA-256 hash shown for each file?

It is a cryptographic fingerprint of the exact bytes you dropped, computed locally in your browser via the WebCrypto API. Two identical files have the same hash; flipping a single bit produces a completely different hash. We display it alongside the other metadata and include it in the JSON export. Practical uses: confirm that two photos are byte-identical, document chain of custody for forensic work, or record in a report exactly which file you analysed without having to keep a copy.

Can I export the metadata as a file?

Yes. Each file card has an "Export JSON" button that downloads a structured .json document containing the file information, every parsed EXIF / IPTC / XMP / ICC tag, the magic bytes, and the SHA-256 hash. The bulk action "Export all metadata (JSON)" bundles every loaded file into a single JSON file. The export is generated entirely in your browser and saved locally — nothing is sent anywhere. Common use cases: feed the JSON into your own analysis script, attach it to a forensic report, or archive the metadata before stripping the originals.

Does it support camera RAW files (CR2, NEF, ARW, DNG)?

Yes, for reading. TIFF-based RAW formats — Canon CR2, Nikon NEF, Sony ARW, Adobe DNG, Olympus ORF, Pentax PEF, Panasonic RW2 — go straight through our TIFF parser, so the full camera EXIF, lens information, dates and GPS appear as expected. Canon CR3 (ISO BMFF, similar to HEIC) and Fujifilm RAF are recognised as view-only. RAW files are not stripped because removing metadata from a RAW would require re-encoding the proprietary sensor data, which the browser cannot safely do. If you need a clean RAW, develop it to JPEG or TIFF first in your editor, then strip with this tool.

Why does the tool show ICC profile details like "Display P3" or "Adobe RGB"?

Many photos carry an embedded color profile that tells viewers how to interpret the colors. We parse that profile locally and surface its actual name (sRGB IEC61966-2.1, Display P3, Adobe RGB (1998), a vendor calibration, etc.) along with color space, device class, version and manufacturer — instead of just showing "[present, N bytes]". This is useful when you want to know whether a photo was tagged with a wide-gamut profile (which can change how it looks on different screens), or to confirm that the cleaned output preserves the profile if you opted in to keep it via the "Keep color profile" toggle.

What are the privacy profiles, and which one should I pick?

The profiles are one-click presets that match common real-world goals so you do not have to know which checkboxes to tick. "Anonymous" removes everything we can — GPS, camera info, dates, author, color profile, orientation tag, and it cleans the filename — best for posting on a forum, on a marketplace or in a public album where you want zero traceability. "Social media safe" is the same but keeps the orientation tag and color profile so the photo still displays the right way up with correct colors on wide-gamut screens — best for posting on Facebook, Instagram, Reddit, Discord. "Keep camera info" removes GPS, dates and author but leaves Make, Model, lens, ISO, aperture and shutter speed alone — useful for portfolio photographers who want to share technical info but not when or where the photo was taken. "Just remove GPS" makes the smallest possible change: only the location is stripped, everything else stays. Pick whichever matches your situation; you can switch profiles per file via the dropdown on each card before clicking "Apply profile".

What is Error Level Analysis (ELA)?

ELA is a forensic technique for spotting edited regions in a JPEG. The tool re-saves your photo at quality 90 and shows you the per-pixel difference between the original and the re-saved version, amplified ~20×. In a genuine, untouched JPEG, the entire image has uniform "error noise" because every region was compressed the same number of times. If someone pasted a face from another photo into this one, that pasted region was compressed differently and shows up as visibly brighter or darker than its surroundings in the ELA result. ELA is heuristic — bright regions can also be sharp edges or strong textures, not just edits — but it's the same first-look tool used by photo-forensics analysts. The whole computation runs locally in your browser via Canvas; your photo is never uploaded.

What does the JPEG Quality estimate mean?

Every JPEG carries a Quantization Table (DQT marker) that defines how aggressively the image was compressed. By comparing your file's DQT against the JPEG-standard reference table for quality 50, the tool reverse-engineers the quality (1-100) the file was last saved at. Quality 95+ is typical for straight-from-camera or lossless export. 80-94 is "saved for web/email" — the most common range. 60-79 means the file was heavily re-compressed (e.g. WhatsApp/Messenger forwarding, 3rd-generation copy). Below 60 is rare except after multiple compression cycles. A photo that claims to be "original from camera" but reports quality 75 is suspicious. Read-only — touches no bytes.

What is the Compression Fingerprint card?

It identifies which software last touched your JPEG by matching the file's quantization tables against twelve known Adobe Photoshop "Save for Web" profiles plus the libjpeg-standard reference, and by inventorying every APP-marker (JFIF, EXIF, ICC, Adobe APP14, Photoshop IRB, MPF, XMP, C2PA, Ducky…). The card combines those signals into a single plain-language verdict like "Almost certainly saved by Photoshop Save for Web at quality 8" or "libjpeg-compatible encoder at quality ≈85", plus a likely pipeline ("Camera capture → Photo editor → Messenger"). All matches are read-only and run locally.

What does the Structure Inspector show?

It walks the file end-to-end and lists every JPEG marker, PNG chunk or HEIC/MP4 box as a clickable row, colour-coded by category (container, metadata, image data, trailing, unknown). Each row shows the marker name, byte offset, length and a short hex preview; clicking expands the full hex dump of that segment. The killer use case is finding polyglot files: data sneaked in after the JPEG EOI or PNG IEND marker — a classic steganography trick — surfaces as a red TRAILING row. Also exposes exotic APP-markers from custom OEM tools or steganography software that other inspectors hide.

What is the JPEG Ghost analysis?

JPEG Ghost is a forensic technique (Hany Farid, 2009) that re-saves the photo at eight quality levels (50, 60, 70, 75, 80, 85, 90, 95) and finds, for every 16×16 block, the quality at which the pixel difference is smallest. In a genuine, untouched JPEG, almost every block will have its minimum at the same Q — the quality the file was originally saved at. If a region was inserted from a JPEG saved at a different quality, that region will have its minimum at a different Q and stand out on the heatmap. The tool flags the inserted area with a red dashed box. Heuristic — sharp edges and textured areas can naturally pick different Qs without being insertions.

What do the Noise and CFA maps tell me?

Two pixel-level checks that run together. The Noise variance map measures how much high-frequency noise each 16×16 block contains; suspiciously smooth blocks light up cyan and indicate possible denoising, beauty-retouching, clone-stamp work or a region painted over from a stock photo. The CFA / Bayer pattern map checks whether the image still carries the camera-sensor demosaicing footprint by comparing the four 2×2 sub-sample positions of the green channel; real photos have a strong Bayer signature (score 50–95/100), while AI-generated, fully re-rendered or aggressively denoised images don't (score under 20/100). The CFA score is a defensible, reproducible alternative to black-box AI detectors.

What is Copy-Move Detection?

It hashes every 16×16 image block (skipping flat sky/wall areas) and finds groups of blocks that look identical and are shifted by the same amount. That's the signature of a clone-stamp tool: when a retoucher paints a person out by sampling nearby grass or sky and pasting it over the person, the painted-over blocks are exact copies of their source blocks, all displaced by the same vector. The heatmap highlights each cluster in a different colour and draws lines between source and destination blocks; a typical "person painted out" shows up as a cluster of 20–100 paired blocks. Runs in a Web Worker so the page stays responsive. False positives are possible on highly repetitive textures (brick walls, fabric patterns) — always inspect the highlighted regions.

What is Thumbnail-Mismatch detection?

Almost every JPEG from a smartphone or DSLR contains a small 160-200px preview image embedded inside the EXIF block — the "IFD1 thumbnail". When somebody edits the photo in Photoshop the main image gets the edits, but the embedded thumbnail is often left untouched. The tool extracts both, computes a 64-bit perceptual hash of each, and reports the Hamming distance. 0-3 bits = identical (no manipulation), 4-15 = match (normal compression difference), 16-30 = suspicious (possible levels/curves/retouch), 31+ = the thumbnail and main image visually disagree, which is a near-bulletproof manipulation proof. Hany Farid published this method in 2009 and it still catches most real-world Photoshop edits.

What does the RGB histogram tell me?

It plots how many pixels each brightness value (0-255) has, separately for the red, green and blue channels. A photograph straight from the sensor has a smooth, continuous histogram — every value from 0 to 255 has at least a few pixels. When somebody applies levels, curves or tone-mapping in Photoshop, the values get redistributed across the 256 bins and certain values become unreachable, leaving characteristic empty bins (gaps). The tool counts those empty bins inside the active brightness range and gives a verdict: under 5 gaps = natural, 5-19 = mild editing, 20+ = strong editing footprint. Native camera output rarely shows more than 5-10 gaps, so a count of 30+ is almost always Photoshop or Lightroom levels.

What is the pixel magnifier?

Hover (or long-press on touch devices) over any forensic heatmap and a small 88×88 pixel zoom-bubble follows your cursor, showing an 8× magnified view of the pixels under the cursor with a yellow crosshair on the centre pixel. The bubble also displays the exact RGB values and hex colour code of that pixel, plus its X/Y coordinates within the image. This is the same workflow Forensically uses and is essential for forensic interpretation: ELA, Noise and CFA heatmaps look very different at 1× than at 8×, and tampering hints often only become visible when you can read individual pixel colours.

Can I export a forensic report as PDF?

Yes. Click "Generate forensic report" in the Pro mode card and a clean print-ready document opens in a new browser tab. It contains the file name, format and size, all cryptographic hashes (SHA-256/SHA-1/SHA-384/SHA-512 plus pHash), key EXIF fields (camera make/model, dates, GPS, software), the compression-fingerprint verdict and every forensic finding you have run (ELA, JPEG-Ghost, Noise/CFA, Copy-Move, Thumbnail-Mismatch, RGB Histogram), each with verdict and heatmap thumbnail. The new tab has a "Print / Save as PDF" button — your browser's standard print dialog produces a PDF you can archive as evidence. Everything is generated in the browser; no upload, no server, no third-party PDF library.

What is the side-by-side compare slider?

Drop two images into the tool, then in the second file's Pro mode card pick the first one from the "Compare with" dropdown — both images are layered on top of each other and a slider in the middle lets you reveal one or the other by dragging. Classic before/after forensic visualization: handy for spotting Photoshop edits when you have an "original" and a "circulated" version of the same photo, for comparing same-scene shots seconds apart, or for spotting compositing differences. The pixel magnifier works on both halves so you can read pixel values across the slider seam. Drop, drop, compare — that's it.

What is the C2PA Self-Sign feature?

C2PA is a new industry standard that lets a photo carry a built-in "digital seal" proving where it came from. Big news agencies (Reuters, AP) and camera makers (Sony A7-IV, Nikon Z9, Leica M11-P) use it to mark their photos as real and unedited; AI tools like DALL-E use it to mark images as AI-generated. With this tool you can put your own seal on any JPEG, generated entirely in your browser. Think of it like signing a letter with your own pen: the seal proves YOU produced this exact version of the file at this exact moment. Useful for: photographers documenting their own workflow, lawyers building evidence chains, journalists sharing photos with editors who know their public key, or anyone testing C2PA tools. How it works: your browser generates a fresh cryptographic key pair, signs the JPEG, embeds the signature into the file, and throws the key away — the signed JPEG is offered for download. Nothing is ever uploaded to us. The recipient can verify the signature at contentcredentials.org. Important limitation: this is "self-signed", which means other tools will display "unknown issuer" — that is expected and correct. For Reuters-grade attestation (where the world recognises you as the issuer) you would need a paid commercial certificate authority. The browser self-sign here is the right level for internal evidence and personal use, and it preserves the tool's strict no-upload privacy promise.

Can I edit individual EXIF fields, e.g. change the camera model or copyright?

Yes — for JPEG photos. Click the "Edit tags" button on a file card to open a form with the most useful editable fields: Make, Model, Software, Artist, Copyright, ImageDescription, DateTime, DateTimeOriginal and Orientation. Whatever you type is written into the EXIF block; leave a field blank to keep its current value. The image pixels are not re-encoded — only the metadata changes. Use cases: scrub a previous photographer's name out of an image you are reusing, set a generic camera maker so the photo cannot be linked back to your specific body, set the date to "1970:01:01 00:00:00" to fully obfuscate timing, or set a copyright notice on photos before sharing. Note: editing PNG, WebP and other formats is not supported yet — for those, use the strip buttons.

Can I replace the real GPS coordinates with fake ones?

Yes, on JPEG. Click "Set fake GPS" on a file card and either pick a famous landmark from the preset dropdown (Eiffel Tower, Statue of Liberty, Brandenburg Gate, Sydney Opera House, Taj Mahal, Great Wall, Machu Picchu, or "Null Island" at 0,0) or type your own latitude and longitude. The fake coordinates replace the real ones in the EXIF block, so anyone who later inspects the photo sees the spoofed location instead of yours. Why not just remove GPS? Because some platforms or AI moderation systems flag photos with no GPS as suspicious, while photos with "any" GPS are treated as normal. Spoofing also lets you keep the data shape intact for archival purposes.

What is "ExifTool-compatible JSON"?

Phil Harvey's ExifTool (https://exiftool.org/) is the de-facto standard command-line metadata tool used in forensics, OSINT and photo workflows since 2003. Its `-j` mode emits a flat JSON array that countless scripts and pipelines already know how to consume. We provide an export option that mimics that format — group-prefixed keys like "EXIF:DateTimeOriginal" or "GPS:GPSLatitude" — so you can drop our exports into any existing ExifTool-aware pipeline without writing a translator. Our coverage is not 100% identical (we do not parse every obscure tag), but for the standard EXIF / IPTC / XMP / GPS fields the output is interchangeable. Switch the export format with the "Export format" dropdown above the file list.

What does "Deep inspection" reveal?

Camera manufacturers reserve a special EXIF tag called MakerNote (tag 0x927C) for their own proprietary data. Each maker uses a different binary format that they document only sparsely (or not at all). When Deep Inspection is off, we report the MakerNote as one opaque blob — "[binary data, 12345 bytes]". When you turn Deep Inspection on, we decode the MakerNote for the three biggest manufacturers (Canon, Nikon, Sony) and surface a curated list of high-value tags: lens model and serial, body internal serial number, shutter count (Nikon DSLRs — useful when buying used), firmware version, picture-control settings, AF info, and many more. This is the same information forensic and OSINT tools use to fingerprint cameras and correlate "anonymous" photos back to a specific device.

Why does the tool change the filename when downloading?

Because filenames themselves leak metadata. Smartphones often save photos as IMG_20240315_142233.jpg or PXL_20240315_142233_AbsolutePath.jpg — that filename alone tells anyone receiving the file the exact second the photo was taken, and on some phones even the GPS-encoded coordinates. Stripping the EXIF block but keeping the filename leaves that channel wide open. Our tool optionally scrubs date-/time-/coordinate-looking patterns out of the downloaded filename — the contents of your custom filename are kept ("vacation-paris-2024.jpg" stays untouched), only the auto-generated camera prefixes get cleaned. You can disable this with the "Clean dates & coordinates from filenames" toggle if you prefer the original filename.

What is the Privacy Risk Score?

It's a 0-100 number plus a traffic-light badge that summarises how identifiable your file is — without you having to read every tag. Each privacy-sensitive piece of data adds points: GPS coordinates +30, embedded thumbnail with its own GPS +15, hardware serial number +20, camera make/model +10, exact capture date/time +10, author/copyright +15, editing software +5, and so on. Below 30 you're "low risk", below 60 "caution", above that "high risk". Click "Why this score?" to see the exact breakdown, so you understand WHICH metadata items pushed you over the threshold. The score is heuristic — actual risk depends on context (who you are, who sees the file) — but it's a fast first read of "is this file dangerous to share publicly as-is?"

What does the "embedded thumbnail" warning mean?

Many cameras and editors save a small preview image (often 160×120 pixels) inside the EXIF block, separate from the main photo. It's called the IFD1 thumbnail. Most photo viewers ignore it, but it persists when you "remove EXIF" with naive tools because they only strip the main IFD0 / EXIF-IFD / GPS-IFD and forget about IFD1. Worse: the thumbnail can have its OWN GPS coordinates that survive the strip — the classic "ghost GPS" leak that has burned more than one journalist. Our tool detects the thumbnail, warns you when it carries its own GPS, and removes it as part of the standard "Remove all" / profile flow. If you only do a selective strip ("Remove only GPS"), the thumbnail stays, so check the warning and use the full strip when in doubt.

What do the "possible inconsistencies" warnings mean?

When the tool inspects a file, it cross-references several timestamps and identity fields that should be consistent in a genuine photo straight from a camera. If they disagree, the file was probably edited, re-saved, or had its metadata partially modified. Examples: DateTimeOriginal says 2024-03-15 14:22 but GPSDateStamp says 2010-08-12 — somebody backdated the photo. Software field is "Adobe Photoshop 2024" but the photo also has Make/Model from a Canon — the file was edited after capture. Camera Make is "Canon" but the MakerNote inside is recognisably Sony — the file was forged or recompressed in a tool that left the wrong vendor blob. These are heuristic flags, not proof of tampering, but they're the same red flags forensic analysts look for. They give you (and the recipient of your file) a fast smell-test for authenticity.

What is the date-timeline view good for?

A typical photo has 3-5 different date fields (DateTime, DateTimeOriginal, DateTimeDigitized, GPSDateStamp, file modification time, sometimes more for video). In a genuine, untouched file they're all within seconds of each other. The timeline plots all of them on a single chronological axis so you can SEE that — or see at a glance that one of them is years off, which is a strong tampering hint. It's the visual companion to the inconsistency-detector. Useful for journalists verifying source photos, eBay buyers checking "as-new" claims, and anyone curious about a file's full history. The view is collapsed by default; click "Timeline view (N dated tags)" on a file card to expand.

What is the ExifTool command snippet for?

For power users: under each file card there's a collapsible "Show equivalent ExifTool command" section that prints the exact command line that would do the same operation in Phil Harvey's ExifTool (https://exiftool.org/) — the de-facto standard CLI metadata tool used in forensics and OSINT since 2003. Two reasons we expose it: (1) Learning. If you want to start scripting metadata operations across thousands of files in a nightly cron job, ExifTool is the right tool, and our snippet shows you the syntax for the operation you're already doing manually here. (2) Verification. If you don't trust our re-implementation, you can copy our snippet, run it locally with ExifTool, and compare byte-for-byte. The snippet updates live when you change the privacy profile in the dropdown, so it always matches what we'd produce.

What are the "?" icons next to each tag for?

Click any "?" icon in the tag list to see a plain-language explanation of what that EXIF tag actually means, why it's sensitive, and what someone could do with it. The names like "GPSLatitudeRef" or "ExifVersion" are not exactly self-explanatory — and many people remove or keep tags without really knowing what they're choosing. The explanations are short (1-3 sentences each), focused on the privacy angle, and available in your selected language. They cover the ~30 most common privacy-relevant tags. We deliberately do NOT pull explanations from a server; everything is bundled in the page so the privacy guarantee ("no server contact") still holds.

Does the tool track what I do with it?

Short answer: no — the tool itself records absolutely nothing about your usage. Which files you drop, which buttons you click, which tags you remove, which privacy profile you pick, which GPS coordinates you spoof, what your photos contain — none of that ever leaves your browser. The strip operations, the tag editor, the GPS spoofer, the JSON / CSV / ExifTool exports, the Risk-Score breakdown — all of it runs locally and is invisible to any server, including ours. Your tool usage is 100% anonymous. What we DO record, like virtually every other website, is an anonymous pageview counter shared across the whole midgardmud.de site: one ping when you land on the page, a follow-up after 60 seconds, and a duration ping when you leave. The pings contain only the URL path you're on, the referrer header your browser sent, your screen size, your browser language, and a randomly generated hit ID. They do NOT contain any file content, any tag names, any actions you took, any cookies, any fingerprint. We do not associate the pings with you across visits and we do not retain raw IPs. If you don't want even that, append ?notrack to the URL once and the tracker stays off across all later visits (a localStorage flag is set). Tools like Brave, Firefox strict mode and uBlock Origin block the analytics endpoint by default. The privacy boundary that actually matters — your files and what you choose to do with them — is fully on your device.