Remove Metadata

What is metadata in image and PDF files?

Metadata is data about data — information that lives inside a file without being part of its visible content. Images often carry surprisingly much: the camera model, the GPS position where the shot was taken, a date, sometimes even the name of the last application that touched the file. PDFs typically include an author, a title, the program that created them, and timestamps.

That isn’t inherently bad — photographers rely on EXIF to sort their work, libraries index PDF metadata. It becomes a problem when you share a file and unknowingly hand over information you didn’t intend to share. Reddit, for example, has a documented history of preserving GPS metadata in uploaded images. Wikipedia documents the privacy implications of EXIF metadata in detail, including incidents where uploaded photos exposed home addresses.

How does the removal actually work?

Each of the four supported formats stores metadata in a different place — and the tool knows them all.

JPG files keep most fields in an APP1 segment right after the SOI marker. The tool walks the segments, identifies EXIF and XMP blocks, and writes only the image-relevant segments back to disk. The orientation tag optionally stays — otherwise your portrait shot would suddenly appear rotated.

PNG files are a chain of chunks; tEXt, iTXt and zTXt are the three text- metadata chunk types. They are filtered out, the remaining chunks are re-assembled with correctly recomputed CRC checksums. WebP works the same way inside its RIFF container; the VP8X flag byte is updated so decoders don’t look for EXIF chunks that are no longer present.

PDFs use a different path: instead of byte filtering, the PDF structure is re-emitted via an open-source library while preserving the cross-reference table and object streams. Document-Info-dictionary fields are reset to empty, the /Metadata stream attached to the catalog is dropped. For signed or encrypted PDFs the tool refuses — a broken signature is worse than a bit of leftover metadata.

How does the tool keep your files private?

Many free online tools for the same job upload your files to their servers, promise “24-hour deletion” and usually deliver on that. But: the upload itself is the issue. Whoever runs the server sees the content — and the server typically isn’t in your jurisdiction. Industry analysts at ISACA flagged EXIF data as a subtle cybersecurity risk in 2025 because of exactly this exposure.

Nothing of that happens here. Once the page is loaded, every step runs in your browser tab via standard web APIs (Canvas, WebAssembly). You can open the Network panel of your developer tools and watch: no file leaves your device.

What this tool honestly does NOT do

We don’t promise “forensically clean” or “untraceable” — that would be dishonest. The tool removes documented metadata fields in JPG, PNG, WebP and PDF at the depth described above. What stays out of scope:

• Steganography: data hidden inside image pixels that requires special tools to detect.
• Pixel watermarks: visible or invisible marks embedded in the image content itself.
• HEIC, MP4, DOCX, XLSX, PPTX: different containers, different walkers — coming.
• PDF annotations, form-field data, embedded files, OCG layers: those live below the Document-Info dictionary and need separate handling.

The tool does what it does, well — and clearly states where its boundary is.

Which related document tools exist?

Other tools from the kittokit ecosystem that fit the topic:

• PDF Password Remover — drop the password requirement on PDFs, locally, no upload.
• Background Remover — extract the foreground subject from an image with neural-network inference in your browser.
• Image Upscaler — upscale images via a neural network in the browser, no server upload.