Drop in a file you already know is broken. This runs identification, structural validation, and near-miss header checks together, classifies what's actually wrong, estimates how much of the file is recoverable where that's a real computable thing — and tells you exactly which tool to open next.
If you're here because something threw one of these, this tool shows you exactly what's wrong and how much is recoverable:
This analyzer classifies and routes — it doesn't repair anything or extract recoverable data itself; that stays a job for the specialist tools it points you to (or, for data carving/thumbnail extraction, isn't built yet at all). The recoverable-byte estimate is only computed where a file format actually has a locatable "this is where it broke" point — PNG chunks, JPEG markers, a short read against a declared size. For formats without that kind of internal structure (most of the 120+ signatures in the database, and PDF's object graph in particular) it says so honestly instead of guessing a number. The classification is a deterministic read of real evidence, not a machine-learned or statistical confidence score.