Start for free
Book Import

Manuscript Ingestion & AI Parsing

Prosewell's Book Import engine lets you import existing draft files into structured Prosewell Books. Powered by background AI workers, it parses chapter boundaries, scene breaks, narrative metadata, and worldbuilding entities in minutes.

Supported File Formats & Text Segmentation

You can import existing manuscripts directly from your local drive. Raw text is ingested and split into contiguous, indexed text blocks for background processing:

Microsoft Word (.docx)

Directly uploads and extracts clean raw text from Word documents while preserving paragraph structure and heading breaks.

Plain Text & Markdown (.txt)

Ingests plain text manuscripts or copy-pasted drafts, automatically segmenting paragraphs into indexable text blocks.

AI Structure & Metadata Discovery

Once uploaded, Prosewell's background AI workers perform top-down semantic parsing over the manuscript text blocks to extrapolate story structure and metadata:

Narrative Metadata

Derives the primary Book Title, Narrative Tense (Past, Present, Future), and Point of View (First Person, Third Person Limited / Omniscient).

Chapter Categorization

Groups blocks into Frontmatter (Copyright, Epigraph), Narrative Chapters, and Backmatter (Acknowledgments, About Author).

Scene Break Detection

Identifies transition markers (such as *** dinkuses or paragraph spacing) to divide chapters into individual scenes.

Worldbuilding Entity Discovery

When initiating a book import, you can choose to Extrapolate Worldbuilding Entities. As the manuscript is parsed, AI scans for recurring story elements and generates Compendium suggestions:

Automated Compendium Bootstrapping

Discovered characters, locations, lore items, and factions are extracted along with reasoning quotes from the manuscript. You can review, approve, or refine these suggestions to instantly populate your Universe Compendium.

Asynchronous Background Ingestion

Book imports are processed entirely in background worker queues without slowing down your browser or locking your screen:

Multi-Stage Pipeline & Live Progress

Jobs progress transparently through three background stages—file upload & text extraction, parallel chapter parsing, and final document assembly—updating live progress indicators from 0% to 100%. Once finalized, your manuscript opens directly in the editor ready for writing.