> ## Documentation Index
> Fetch the complete documentation index at: https://docs.quarkle.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# How to Get AI Feedback on Your Manuscript

> A step-by-step workflow for revising a finished draft: AI-generated manuscript reports, line-level editing suggestions, and chat-based analysis.

You've got pages already, and you want to know what's working. This guide is the order that gets you there fastest: sync, zoom out, then zoom in.

## Make sure the Atlas is caught up

Chat and the Review Room both draw on the Story Atlas: Chat broadly, the Review Room through character profiles and the current chapter's summary. If you've just imported a draft or made a big pass of edits, give it a moment to catch up before you lean on either. It updates itself as you write, or press **sync** to force it.

<Frame caption="The Story Atlas, synced and ready.">
  <img src="https://mintcdn.com/quarkle/HrAp7UTT41GKfNkT/images/app/atlas-files.jpg?fit=max&auto=format&n=HrAp7UTT41GKfNkT&q=85&s=de4dd6d1c615703a2c614ef65783f531" alt="The Story Atlas file tree with System notes and auto-discovered Character and World notes" width="1568" height="721" data-path="images/app/atlas-files.jpg" />
</Frame>

<Note>
  Reports are the exception: they read your prose directly, not the Atlas, so a stale Atlas won't affect a report's quality. If you only need a quick read on where things stand, you can run one before syncing anything. See [Keeping the Atlas current](/under-the-hood/keeping-the-atlas-current) for how the sync actually works.
</Note>

## Get the big picture with a Report

Before you dive into line edits, step back. Open the **Report** tab, pick the **Editorial letter** tier, set the scope to **Whole book**, and send it to Scribe. It reads the way a first reader would and writes up a verdict, what's working, and a revision plan sorted by how much each fix actually matters.

<Frame caption="Pick a tier and a scope, then send it to Scribe.">
  <img src="https://mintcdn.com/quarkle/3Z-v7kzGlib5E3ID/images/app/reports-config.jpg?fit=max&auto=format&n=3Z-v7kzGlib5E3ID&q=85&s=1251fea2f3abe74ebe2b9032c86f9a73" alt="The Reports Config tab showing the four tier options, scope picker, and send to Scribe button" width="1036" height="1155" data-path="images/app/reports-config.jpg" />
</Frame>

This tells you *where* to spend your editing time before you commit to a line-by-line pass. Full details: [Reports](/ai/reports).

## Run the Review Room for line-level notes

With a sense of the big picture, open **Review** and run your reviewers against the chapters that need the most work. Turn on the tiers that matter right now (skip grammar reviewers if you're deep in a structural pass) and click **Run review**.

<Frame caption="The Review Room. Reviewers are grouped into tiers. Turn them on or off, then run a review.">
  <img src="https://mintcdn.com/quarkle/pvbq7pupzmmMGYpd/images/app/review-room.jpg?fit=max&auto=format&n=pvbq7pupzmmMGYpd&q=85&s=c2acd167c559273a8311034adb137c10" alt="The Review Room showing reviewers grouped into tiers with comment counts" width="1568" height="721" data-path="images/app/review-room.jpg" />
</Frame>

Comments land as beads in the margin, each with a suggested rewrite previewed as a diff.

<Frame caption="Click a bead to see the note and its rewrite, previewed inline as a diff.">
  <img src="https://mintcdn.com/quarkle/5WUUaYl-dgOEAFoM/images/app/review-comment-diff.png?fit=max&auto=format&n=5WUUaYl-dgOEAFoM&q=85&s=58eb79124037beb69708dec5303e9f6b" alt="A comment open in the manuscript, showing a suggested rewrite as an inline diff with accept and reject controls" width="944" height="384" data-path="images/app/review-comment-diff.png" />
</Frame>

Work through them one at a time: **Accept** the rewrites that land, **Reject** the ones that don't, **Expand** any note you want the full reasoning on. Full walkthrough: [The Review Room](/ai/review-room).

## Ask Chat about specific concerns

Some questions don't fit a bead in the margin. *"Does Mara's motivation in chapter 6 track with what we established in chapter 2?"* or *"is the pacing too slow through the middle act?"* are Chat questions, not review comments. Chat reads the relevant chapters and notes before answering, and can point you to the exact lines it's talking about.

<Frame caption="Chat answering a question about the manuscript, with a trail of what it read.">
  <img src="https://mintcdn.com/quarkle/5WUUaYl-dgOEAFoM/images/app/chat-response.png?fit=max&auto=format&n=5WUUaYl-dgOEAFoM&q=85&s=f0a68a739c655528778f5a1d3d7b8486" alt="A Quarkle Chat exchange with a tool trail showing Notes and Search" width="1066" height="850" data-path="images/app/chat-response.png" />
</Frame>

## Work through it methodically

A practical order that keeps you from getting lost in a long manuscript:

<Steps>
  <Step title="Report first">
    Generate a whole-book report. Note the two or three highest-leverage suggestions.
  </Step>

  <Step title="Review the flagged chapters">
    Run the Review Room on the chapters the report called out, not the whole book at once.
  </Step>

  <Step title="Ask Chat about anything structural">
    Continuity, pacing, and character-consistency questions that a line-level comment can't capture.
  </Step>

  <Step title="Re-run the report once you're done">
    A fresh whole-book report after a revision pass tells you whether the changes actually moved the needle.
  </Step>
</Steps>

<Tip>
  Reports are capped at 10 a day and need at least 500 words to run; reviews need at least 100 words and one reviewer on. Pace a long revision pass across a few sessions rather than burning through your report budget in one sitting.
</Tip>

<Card title="Next: Mastering the Story Atlas" icon="arrow-right" horizontal href="/guides/mastering-the-story-atlas">
  Get the most out of the memory that powers everything above.
</Card>
