Every established publisher owns a goldmine it can't spend: the backlist. Titles that went out of print before digital typesetting exist only as physical books — no InDesign files, no Word manuscripts, sometimes not even the original typescript. Reissuing them as ebooks or print-on-demand editions means getting the text back out of the printed page.
The same problem hits editorial workflows from the other side: authors deliver typescripts, agents send photocopied sample chapters, and estates hand over boxes of paper. Before any editing can start, someone has to turn paper into text.
The Backlist Conversion Problem
Retyping a 90,000-word novel costs weeks of labour and introduces fresh errors. Naive OCR is faster but produces its own mess: broken hyphenation at every line ending, running heads and page numbers embedded mid-sentence, chapter openings mangled by drop caps, and em-dashes flattened into hyphens that an editor then has to hunt down one by one.
Inkscribe AI's pipeline addresses each of these at the processing stage, not in cleanup:
- Line-break hyphens are rejoined using dictionary and context checks — 'pub-lisher' becomes 'publisher', while genuine compounds keep their hyphens.
- Running heads, folios, and page furniture are detected by position and repetition, and stripped from the text flow.
- Paragraph structure survives: indents and spacing map to real paragraph breaks, not one line per printed line.
- Typographic characters are preserved — em-dashes, curly quotes, ellipses, and accented characters come through as themselves.
- Chapter and section headings are tagged, giving you a structured file ready for ebook conversion.
A Realistic Backlist Workflow
- Scan or photograph the printed book — a flatbed scan at 300 DPI, or the mobile app's camera scanner for quick capture.
- Upload the full book as one batch; a 320-page novel processes in a few minutes.
- Export to Word or Markdown with structure intact — headings, paragraphs, italics.
- Run your normal copyedit pass. Editors report proofreading OCR output at 97–99% accuracy is 5–10x faster than checking a retyped manuscript.
- Feed the clean file into your ebook conversion or typesetting pipeline.
For Editors: Typescripts and Proofs
Editorial teams use the same pipeline on incoming material: a photocopied typescript from an estate, an author's annotated page proofs, a competition's paper submissions. Handwriting recognition picks up marginal corrections, and ScribIQ can summarise a 400-page submission or find every scene featuring a given character — useful triage before a full read.
Rights Departments
ScribIQ also reads contracts. Upload a scanned head contract and ask directly: 'What territory rights were granted?' or 'Is there an audiobook clause?' — answers cite the exact clause on the exact page.
What It Costs
A 320-page book is 320 credits. On the Pro plan that's well within a month's allowance — meaning a small press can digitize a backlist title a month for $12, versus hundreds of dollars for professional rekeying. Larger programmes — university presses, publishers converting whole catalogues — should talk to us about Inkscribe AI Enterprise for unlimited batch processing.
The backlist is the cheapest acquisitions programme a publisher has. The titles are already yours — the text just needs to come back off the page.