A local-first creator toolbox
Choose images or videos · up to 50 files
Remove hidden metadata, GPS location, and C2PA from photos and videos locally. EXIF is camera tags. GPS is coordinates stored in the file, when they were written. Supported C2PA Content Credentials are signed origin records — a missing packet does not prove a photo is human-made. Up to 50 files. No account. No upload. Just your browser.
C2PA: Coalition for Content Provenance and Authenticity; C2PA specification 2.4
Last updated
Learn what may be inside a file: EXIF metadata · GPS metadata · C2PA
EXIF is camera tags. GPS is coordinates stored in the file, when they were written. Supported C2PA Content Credentials are signed origin records — a missing packet does not prove a photo is human-made.
The original file stays. This tab writes a copy and removes supported metadata from that copy: EXIF, GPS, XMP, and supported C2PA. Up to 50 files. That clean is not uploaded.
Drop more files onto this strip
Total files
0 Items
| File | Type / Size | Risk assessment | Status | Action |
|---|
Ready to sanitize 0 files
Choose
Your originals stay untouched
Inspect
See the hidden information
Clean
Remove supported metadata
Save
Take a cleaner copy with you
The toolkit / 24
Remove location, camera details and embedded AI-origin info before you share. Supports C2PA, EXIF, GPS, XMP and selected MP4/MOV metadata.
Read EXIF, GPS, XMP, and IPTC from a JPEG, PNG, or WebP in this browser.
Local inspection, then optional official OpenAI C2PA / SynthID.
Remote check for AI-written or AI-assisted text. Not a local-only tool.
Hardware-first WebCodecs encoding in your browser. No upload, no server queue. WebCodecs is required; there is no WASM fallback.
Shift SRT or WebVTT cues, or stretch them to fix drift, in this browser. Nothing is uploaded.
Read MP4 and MOV container tags locally, including creation time and location tags when the file stores them.
Scale 1080p video up to 4K or 2× client-side with Rust WebAssembly and SIMD. No server queue, no cloud inference, no hallucinated pixels.
Make a smaller clip for Discord’s free upload size in this browser. The file stays on your device. The starting target sits under the published non-Nitro cap, and that is not a promise Discord will accept the file.
Make a smaller copy for X in this browser. The size box starts at the published non-subscriber file maximum. Duration is a separate rule.
Make a smaller Bluesky copy in this browser. The 2024-09-11 post allows one video up to 60 seconds and a daily total. The 2026-01-26 sentence about three-minute videos is not a file-size cap. The size box stays empty.
Make a smaller Telegram copy in this browser. The FAQ says 2 GB each, or 4 GB with Premium. This page does not turn that figure into a megabyte target.
GIF to Live Photo: an iPhone JPEG+MOV pair, or a Samsung / Android Motion Photo file. Encoding stays in this browser. Phone import is Beta.
Extract text from images and scanned PDFs locally, with English, 简体中文, 繁體中文 and 日本語 support. Export TXT, Markdown or searchable PDF.
Merge, split, reorder, rotate, extract and convert PDFs locally. Structural operations copy PDF pages instead of flattening them to images.
View PDF metadata: author, creator app, dates, and XMP when the packet can be read.
Scan, redact, summarize, rewrite and extract from pasted text in the browser. Privacy Scan works without downloading a model. Optional MiniCPM5-2B runs locally on WebGPU.
See which browser signals this tab can expose. Nothing is scanned until you start it. A blocked API is unavailable, not proof you are anonymous.
Compare a local digest of canvas, WebGL, fonts and audio signals in this tab. The digest stays here and is not an ID of you.
Draw the same canvas twice in this tab and hash the full pixel buffer. A change shows the result moved; it does not prove you are unique or anonymous.
List ICE candidates after you start the test. An mDNS name is not a private IP. A different public address is an observation, not proof a VPN failed.
Convert a HEIC or HEIF photo to JPEG or PNG in this browser, then drop EXIF and GPS from that new image. The homepage cleaner does not open HEIC.
Replace names you list, email addresses, and common API-key shapes in this browser before a ChatGPT paste. The text is not uploaded.
Draft a Telegram account-recovery appeal: what happened, what you want, and a message you can send. It does not submit the appeal for you.
Inspect and remove metadata from photos in this browser tab — EXIF, GPS location, XMP, and supported C2PA Content Credentials. No upload. Invisible pixel watermarks are not metadata and are not removed.
PrivacyThe sharing copy is rebuilt in this tab. Supported image headers are dropped by Canvas re-encode; supported video boxes are removed without transcoding samples.
JPG, PNG, WebP, MP4, and MOV. Up to 50 files in this browser. Press F12, then Network, and confirm the file is not posted. Supported C2PA metadata can be viewed and removed from the sharing copy. Pixel watermarks are not metadata and may remain.
This workflow processes your file in this browser tab. Opening DevTools is optional; it is the simplest way to watch what the page actually requests.
Expected Network activity: this site’s CSS/JS, and — after you run the tool — same-origin WASM, codec, or OCR language data. That is a download of our engine, not an upload of your file.
This panel is a product fact, not a live packet sniffer. We do not invent a request counter. Ads, extensions, and your own later clicks can still create network traffic.
Platform walkthroughs for C2PA labels, EXIF GPS leaks, and client-side sanitization.