Editorial Analysis
Editorial Analysis is the first tab in the Analysis workspace, and it’s the one you’ll open most often. It answers three distinct questions about your writing in three distinct phases, and the phases exist in a specific order because the order matches how you should actually diagnose writing problems. You start with rhythm, because if your output cadence is broken then no amount of line-editing is going to save you. You move to line pressure, because if your rhythm is fine but your prose is still failing, the problem is probably craft-level. And you finish with scene shape, because if your rhythm and your prose are both solid and you’re still unhappy, the problem is structural — and structural problems are the hardest to diagnose without stepping back from the sentence level.
The three panels in Editorial Analysis are Writing Stats, ProseGuard, and Manuscript Review. They share a sidebar rail in the Editorial tab, and the rail is ordered left-to-right the way you should usually work them — rhythm first, pressure second, shape third.
Phase 1 — Writing Stats (the rhythm question)
Section titled “Phase 1 — Writing Stats (the rhythm question)”Question: Am I writing enough? Am I writing consistently? Is this a time problem or a craft problem?
Panel: The Writing Stats view in the Editorial sidebar rail.
Writing Stats is the cadence panel. It shows your output over time — words per day, per week, per month — alongside session data, writing streaks, and goal progress. The default view is the last 30 days, with your current streak highlighted and your session-based goals visible.
This panel exists because before you can fix a writing problem, you have to know whether the problem is actually a writing problem at all. An author who feels like their prose is weak often discovers, when they look at the Writing Stats panel, that they haven’t written anything in four days — and the weakness they feel isn’t prose weakness, it’s rust. The fix for that isn’t editing harder. It’s writing more.
Most authors underestimate how often their craft problem is actually a cadence problem. Writing Stats is the panel that surfaces the cadence data so you can tell the difference.
What to do with the answer. If your cadence is broken, the fix is writing more, not editing harder. Move to the editor, open the document you’re avoiding, and put down words. If your cadence is fine — you’ve been writing every day, you’re hitting your word counts, your streak is intact — then the problem isn’t rhythm and you move to Phase 2.
Phase 2 — ProseGuard (the line pressure question)
Section titled “Phase 2 — ProseGuard (the line pressure question)”Question: Is my prose quality holding up? Am I breaking my own rules? Am I drifting into machine-like patterns without noticing?
Panel: The ProseGuard view in the Editorial sidebar rail.
ProseGuard is the deterministic prose linter. It reads the document you’re looking at against a set of rules you’ve configured — banned phrases, banned words, flagged Machine Tells, filter words, POV consistency rules, character pronoun requirements, forbidden terms, required name forms, em-dash limits, ellipsis limits — and produces a structured list of violations with severities.
The full ProseGuard documentation covers every rule category and how to configure them. What matters for the Editorial Analysis workflow is that this is the second panel you open when the rhythm is fine but something still feels off. ProseGuard tells you which sentences are breaking which rules, at the character-offset level, with severities and suggested fixes.
This is the pressure phase — you’re under editorial pressure at the sentence level, and ProseGuard is how you find the sentences that need work.
What to do with the answer. If ProseGuard flags a lot of issues, work through them in the editor. Violations appear as inline highlights in the document, so you can fix them in-place. Once you’ve resolved the critical and major issues, re-run the check and see how the count has dropped. When ProseGuard is mostly clean but the chapter still feels off, you move to Phase 3.
Phase 3 — Manuscript Review (the scene shape question)
Section titled “Phase 3 — Manuscript Review (the scene shape question)”Question: Is the chapter actually working at the scene level? Are the scenes paced right? Is the POV balance correct? Are the scene weights appropriate?
Panel: The Manuscript Review view in the Editorial sidebar rail.
Manuscript Review is the structural phase. It reads the whole manuscript (or a chapter) and produces scene-level analysis: pacing curve across the chapter, POV distribution, scene weights, dialog-to-action ratios, character presence per scene. The output is visual — charts and heatmaps — because structural problems are usually easier to see than to describe.
A chapter where one character dominates 80% of the scene weight but the plot supposedly has three main characters is a structural problem, not a line-level problem. ProseGuard won’t catch it (every sentence is technically fine). Writing Stats won’t catch it (you wrote the chapter on schedule). Manuscript Review is the panel that makes it obvious.
This is the phase most authors skip, which is a mistake. A lot of the “I don’t know why this chapter isn’t working” problems are structural, and structural problems look invisible at the sentence level. You can have every sentence ProseGuard-clean, every day meeting your word count, and still have a chapter where the protagonist disappears for 2,500 words in the middle. Manuscript Review surfaces it.
What to do with the answer. Structural fixes are bigger fixes. You’re not tweaking sentences — you’re re-ordering scenes, cutting subplots, adjusting POV, rebalancing character presence. These changes take longer than a ProseGuard pass and they’re usually the kind of thing a developmental editor would flag. Manuscript Review is a first-pass check that catches the most obvious structural issues before you send the chapter to a real editor.
Why the order matters
Section titled “Why the order matters”The three phases are ordered because each one catches a different kind of problem, and diagnosing them in the wrong order wastes effort.
If you skip Phase 1 and start editing sentences, you might spend an hour polishing prose when the real problem is that you haven’t written enough new material in a week. You’ll finish the editing session feeling productive but not actually having moved the project forward.
If you skip Phase 2 and jump to Manuscript Review, you might start rewriting scene structure when the real problem is a few machine-flavored phrases slipping through the prose. You’ll restructure a chapter that was actually fine and waste time solving the wrong problem.
Running the phases in order catches the cheapest-to-diagnose problems first. Writing Stats is an instant read. ProseGuard is a one-minute scan. Manuscript Review takes longer and produces a heavier analysis. The order matches the cost and the likelihood — cheap, common problems first; expensive, rarer problems second.
How the panels relate to the rest of Ishvana
Section titled “How the panels relate to the rest of Ishvana”Editorial Analysis isn’t a self-contained feature. It’s a view on top of data and rules that live elsewhere.
- Writing Stats reads session data from the analytics system, which is always running in the background when you have “Track writing sessions & goals” enabled in Settings.
- ProseGuard reads the rules you’ve configured for your project (plus document-level, scene-level, and character-level overrides) and runs them against the document text. The rules live in the ProseGuard module; the Editorial panel just runs the checker and displays results.
- Manuscript Review reads your outline structure, your document contents, and the entity mentions detected by Entity Detection. It produces visualizations from that data.
All three panels update as you work. Run a ProseGuard check after an editing pass and the results update. Write new words and Writing Stats updates on the next refresh. Restructure a chapter and Manuscript Review reflects the new shape the next time you open it.