mechanical-preprocess

Standardize em-dashes and remove forbidden words across a multi-chapter manuscript by running a bash-driven preprocessing pipeline that performs deterministic pattern matching and safe text replacements before any AI editing.

Cleaning up repetitive phrasing before AI editing is best handled by a mechanical preprocessing workflow that scans chapters, applies safe replacements, and generates a review report with diffs to preserve the author's voice.

Yes, you can use bash scripts for mechanical manuscript cleanup on 30 or more chapters, as the workflow uses bash-first pattern matching and counting to apply safe deterministic transformations across large multi-chapter files.

A phase-based mechanical cleanup workflow handles manuscript text processing through four distinct stages: scanning the text, replacing target patterns, reviewing quality, and reporting with per-chapter diffs and a final summary report.

You should run mechanical text cleanup at the beginning of your manuscript editing workflow, specifically before AI agents touch the text, to reduce mechanical noise and ensure standardized punctuation for later dialogue polishing.

Mechanical manuscript preprocessing preserves your original writing voice by using deterministic bash-driven transformations for punctuation and repetitive structures, which provides safe replacements and outputs diffs for manual review rather than altering style.