Media & Metadata Extraction

An image, an audio file or a subtitle track carries several separable things: the content itself, the data describing it, and the file it is stored inside. These tools read all of them — the words pictured in an image, the camera and location recorded alongside a photo, the colours it is made of, the artwork embedded in a track, and the transcript inside a subtitle file. Everything is decoded in your browser.

Which tool do you need?

If you…Use
You want the words shown in a pictureImage to Text (OCR)
You want the camera, date and location recorded in a photoEXIF Data Extractor
You want the colours an image is made ofColor Palette Extractor
You want the tags and cover art inside an MP3MP3 Tag & Album Art Extractor
You want the transcript out of a .srt or .vtt subtitle fileSRT Subtitle Viewer

If your image is inside a document, take the original out first with the document image tools rather than screenshotting it — a re-compressed copy shifts the colours and drops the metadata.

Content, metadata, and the container

These tools answer three different questions about the same kind of file, and it is worth being clear which one you are asking. Content is what the file depicts: the text OCR reads, the colours a palette measures. Metadata is what the file records about itself: the camera that took a photo, the artist of a track. The container is what else is stored alongside: the cover image inside an MP3, the photographs inside a Word document.

Extraction from a container is lossless, because the thing you want is stored whole and unmodified — an MP3's cover art is a complete JPEG, and a picture in a .docx is the file the author inserted. Reading content is the opposite: OCR infers characters from shapes and a palette summarises millions of pixels into a handful of values. Exactness is available in one case and not the other.

Metadata deserves particular care. A photo's EXIF block frequently records exactly where it was taken, which is why the EXIF extractor shows coordinates but never sends them to a mapping service.

Every media & metadata tool

Frequently asked questions

Can I get the text out of a photo?

Yes — that is OCR. Use image to text. Accuracy depends almost entirely on how sharp and how large the text is in the picture.

Does reading a photo's location upload it anywhere?

No. The photo is parsed in your browser, and the coordinates are shown to you without being turned into a map link, because loading a map would disclose the location to a third party.

Why do the colours I extract differ slightly from the brand's official values?

JPEG compression and colour-profile conversion both shift values a little. Extract from a PNG or the original asset when you need to be exact.

Other extraction groups