Damaged artifacts that contain text make up an “invisible library” of written material that is incredibly difficult to read.  But progress over the past decade using new computer techniques for the digitization and analysis of text found in cultural objects (inscriptions, manuscripts, scrolls) has led to workable, non-invasive methods for reading this invisible library.  This talk shows results over the past two decades from digital restoration projects on Homeric manuscripts, Herculaneum material, and Dead Sea scrolls, culminating in the reading of the text from within a damaged scroll unearthed at En-Gedi, which has been hailed as one of the most significant discoveries in biblical archaeology of the past decade.  Premised on “virtual unwrapping” as an engine for discovery, this presentation explains the complete process developed for reading the scroll from En-Gedi, and the broader significance of the discovery.  The talk concludes by unveiling a new approach – Reference-Amplified Computed Tomography (RACT) – where machine learning becomes a crucial part of the digital restoration pipeline. You will leave this talk considering that RACT may indeed be the pathway for rescuing still-readable text from some of the most stubbornly damaged materials, like the enigmatic Herculaneum scrolls.