Prompt Engineering Context Analysis

Split a Book for AI Analysis — Whole-Work Reading in Parts

Book-length analysis needs the WHOLE work read before any judgment. Analysis mode delivers the chapters in order and triggers the full analysis only at the end.

Overview

Long-form analysis fails when the model judges from the first installment: themes "concluded" from chapter one, character arcs assessed before they arc. Analysis mode is built for whole-work reading — every chunk forbids early analysis, the parts arrive in narrative order with acknowledgements, and the final chunk both declares delivery complete and carries the analysis task, with the instruction to draw on every chunk and cite which chunk evidence comes from. This setup loads book-scale prose with a literary analysis task delivered at exactly the right moment: after the whole text.

Workflow

  1. Deliver in narrative order

    Chapters arrive as ordered chunks — the reading order is part of the analysis contract.

  2. Forbid early conclusions

    Every chunk repeats the wait rule; the analysis cannot begin at chapter one.

  3. Demand chunk citations

    The final task requires evidence to name its chunk — claims become checkable against the text.

Why This Works

  • Whole-work gating matches how long-form meaning actually accumulates
  • Ordered delivery preserves the structure the analysis is about
  • Chunk citations keep the analysis anchored to the actual text

Best for

  • Authors and editors working with full manuscripts
  • Whole-work questions: arcs, structure, recurring imagery
  • Anyone whose long-text analysis came back half-read

Not for

  • A human-readable summary of the book — that's the Structured Summary Prompt
  • Checking whether the book fits at all — the Estimator's plan-large-document territory

Use cases

  • Analyzing manuscripts and long-form drafts as wholes
  • Preventing theme conclusions from chapter one
  • Getting chunk-cited evidence in literary analysis

Tip: Save time by exploring related resources and tools that integrate with this workflow.

Explore all resources