How to Use LoreVia During a Real Revision Week

A revision week can disappear without producing a revision. You open the manuscript intending to repair the middle, reread the first three chapters, adjust several sentences, search for a name, follow an old note into a deleted subplot, and finish the day with a longer list than you began. The problem is not laziness. “Revise the book” is too large to guide a working session.

LoreVia is most useful when it reduces that uncertainty to a sequence: orient, question, inspect, decide, edit, and verify. It should not become another dashboard you browse instead of writing. Give each day one editorial purpose and carry accepted decisions back to the manuscript.

This five-day model assumes that a draft has already been uploaded and processed. Adapt the pace to your life. One “day” can be a two-hour session, and one revision week can stretch across a month. Preserve the order even when the calendar changes.

Choose one revision question before Monday

Begin with a problem that can produce visible decisions. “Make the book better” is not a revision question. Neither is “check the characters.” Useful questions name a pressure, an area of the manuscript, and the consequence you care about.

Examples include:

• Does Mira’s distrust of the council escalate clearly between Chapters 7 and 15?
• Is there enough time for the northern journey before the coronation?
• Which setup promises a confrontation with Dev, and where does that promise pay off?
• Does the reader understand the cost of the gate ritual before the climax uses it?
• Which scenes still assume the removed brother subplot exists?

Choose one primary question and, at most, one secondary check. A narrow week creates a finished slice of revision rather than a tour through every available analysis view.

Write a definition of done. For example: “I will identify every scene that changes Mira’s trust, decide whether each turn stays or changes, edit the affected scenes, and verify that the new sequence has no unsupported jump.” This definition prevents interesting side findings from taking control.

Monday: verify the manuscript model

Monday is orientation, not rewriting. Confirm that the uploaded version is the version you intend to revise. Review chapter boundaries and titles. Scan major character records for duplicates, aliases treated as separate people, or obvious role errors. Check whether the relationships and timeline signals relevant to the week resemble the draft you know.

Do not attempt to perfect every extracted record. Correct errors that would distort the chosen question. If your week concerns Mira and the council, merge Mira’s nickname, verify the council members, and ignore a duplicated shopkeeper unless that person affects the same scenes.

Then establish a baseline. Find the relevant characters, relationships, events, or thread. Note the chapters already surfaced and any expected chapter that appears missing. The baseline is not a verdict; it is a map for inspection.

For a relationship question, record the current sequence of meaningful turns. For a timeline question, list the hard anchors and uncertain intervals. For a dropped-thread question, list the setup, intermediate reminders, and expected payoff. Keep this baseline short enough to compare after editing.

Monday should end with a scene list and two or three sharper questions, not polished prose.

Tuesday: ask for evidence, not evaluation

Use grounded questions to locate passages. Avoid prompts such as “Is the relationship good?” or “Does the timeline work?” Those invite broad interpretation. Ask questions whose answers can be checked against chapters.

For the trust example, useful prompts might be:

• In which chapters does Mira accept or reject help from a council member?
• What does Mira know about the council before Chapter 10?
• Which scenes show the council keeping information from her?
• Where does Mira defend the council despite her stated distrust?

Read the returned passages. Open the surrounding scene because a short excerpt may omit irony, point of view, or the consequence in the next paragraph. Add relevant chapters that the answer missed. Remove passages that share vocabulary without serving the question.

Grounded answers save retrieval time; they do not replace reading. Your job on Tuesday is to build the smallest reliable evidence set from which Wednesday’s decisions can be made.

Wednesday: turn findings into decisions

Wednesday is the editorial centre of the week. Review the evidence in manuscript order. For each scene, decide what function it performs now and what function it must perform in the revised sequence.

Use a simple decision vocabulary:

• Keep: the scene already supports the intended progression.
• Clarify: the scene contains the right turn but the signal is too weak or ambiguous.
• Move: the turn works but appears in the wrong order.
• Rewrite: the scene’s cause, choice, or consequence no longer fits.
• Remove: the beat duplicates another scene or belongs to an abandoned version.

Do not edit while still deciding the whole sequence. A local improvement can make a later scene redundant, and an elegant line can make you reluctant to choose the stronger structural fix. Complete the decision pass first.

For each change, name the downstream scenes that rely on the old state. If Mira begins trusting the council earlier, later suspicion needs a new trigger or softer language. If the journey gains two days, references to fatigue, weather, deadlines, and messages may change. If a setup is removed, search for echoes and payoffs.

End Wednesday with a bounded edit list ordered by chapter. The list should explain the story logic, not prescribe every sentence.

Thursday: revise in the manuscript

Return to the authoritative writing file. Work through the decisions in order. Keep LoreVia open for evidence, but make prose changes where the manuscript lives.

Start with causes before consequences. If Chapter 8 must establish a stronger betrayal, edit it before adjusting the anger in Chapter 12. If a rule requires a cost, establish or clarify the cost before rewriting the climax. This reduces the chance of inventing reactions to an event that later changes again.

After each scene, leave a short implementation note: what changed, what remains uncertain, and which later scenes should be checked. Avoid turning the note into a summary of the whole chapter. It exists to support Friday’s verification.

Protect scope. Line-level opportunities will appear while you revise. Fix a sentence when the new logic requires it; otherwise mark it for a prose pass. Structural and continuity work becomes exhausting when every paragraph is also asked to become beautiful.

If a decision proves wrong in execution, update the plan. The Wednesday list is a hypothesis, not a contract. Record why the change failed so Friday does not compare the new draft to an obsolete intention.

Friday: verify the changed story

Verification is not rereading the same report. The manuscript has changed, so the question must be asked again against the new version. Upload or reprocess the revised material according to the current workflow, then revisit the narrow baseline established on Monday.

Check three layers.

Local coherence: Does each edited scene now contain the intended cause, choice, or clarification? Read enough context to judge tone and point of view.

Sequence coherence: Do the relevant turns, facts, or events progress in a plausible order? Has moving one beat created a new repetition or gap?

Collateral continuity: Did the edit change knowledge, objects, injuries, timeline, relationships, or open promises outside the primary question?

Run direct queries. Search new wording and old wording. Confirm that removed facts no longer appear in downstream assumptions. Inspect chapter citations rather than accepting a clean summary.

Revision is complete when the new manuscript supports the decision, not when the task card says “done.”

A worked example: repairing a reveal

Suppose a thriller reveals in Chapter 18 that Ilya forged the harbour manifest. Beta readers report that the reveal feels obvious and confusing at the same time: obvious because Ilya behaves suspiciously, confusing because no one has established that he could access the records.

On Monday, the author verifies Ilya’s aliases and gathers his appearances. The thread view shows several harbour references but no explicit access. On Tuesday, grounded questions locate scenes where Ilya knows ship schedules, borrows a clerk’s coat, and lies about visiting the customs office. The clues imply interest but not opportunity.

On Wednesday, the author decides to keep the schedule clue, clarify the borrowed coat as access rather than atmosphere, remove one overly suspicious reaction, and add a consequence after the reveal showing why the forgery altered the investigation.

On Thursday, those scenes are revised in cause-first order. On Friday, the author checks who sees the coat, who knows Ilya’s former job, and whether the added consequence conflicts with the timeline. The result is not merely “less confusing.” It is a reveal with a traceable opportunity, balanced suspicion, and visible consequence.

The example works because the week focused on one question. A general request to “analyse the plot” might have produced more observations and fewer edits.

Use each LoreVia view for a specific job

The character views are useful for appearances, aliases, roles, traits, and state across chapters. Relationship views help locate changes between people rather than treating profiles in isolation. Timeline views help arrange events and inspect duration or overlap. Thread and continuity views help surface setups, unresolved obligations, and possible contradictions. Grounded questions help retrieve chapter evidence. Revision planning helps turn accepted findings into work.

Do not visit every view as a ritual. Choose the view that fits the question. If the problem is a relationship turn, begin with the relationship and supporting passages. If the problem is travel, begin with events and anchors. Additional views are useful only when the first investigation reveals a dependency.

Also distinguish extraction correction from manuscript revision. Merging two duplicate character records improves the model. Rewriting the chapter where their alias is introduced changes the book. Both can matter, but they are different tasks with different definitions of done.

Keep the weekly record small

At the end of the week, preserve five things:

1. The revision question.
2. The evidence set or chapter list.
3. The decisions made and why.
4. The manuscript version containing the edits.
5. The verification result and any parked issues.

This creates an editorial history without writing a second manuscript. It also makes future reversals easier. If a later revision reopens the same relationship problem, you can see which assumption guided the earlier fix.

Avoid recording every prompt and interface action. Preserve conclusions and evidence, not a transcript of tool use.

Adapt the rhythm to different kinds of revision

For a continuity week, Monday maps the relevant canon, Tuesday traces conflicts, Wednesday chooses which fact remains true, Thursday propagates the fix, and Friday searches for old assumptions.

For a character week, Monday maps appearances and relationships, Tuesday gathers pressure points, Wednesday defines the intended arc, Thursday revises turning scenes, and Friday checks emotional progression.

For a beta-reader week, Monday groups feedback, Tuesday locates the passages behind repeated comments, Wednesday decides which comments identify the same root problem, Thursday revises, and Friday prepares a cleaner packet or next-round questions.

The shape remains consistent because good revision separates retrieval, judgment, implementation, and verification.

Finish with the next narrow question

Friday’s collateral check will reveal more work. Choose the next question from that list, but do not automatically begin it. Archive the completed week, make a backup, and give yourself enough distance to see whether the chosen fix actually improves the reading experience.

LoreVia should make a revision week more concrete, not more crowded. Use it to find the scene, preserve the evidence, and expose dependencies. Then close the analysis long enough to make the difficult creative choice in the prose.

The rhythm is deliberately modest: one real problem, one evidence set, one group of decisions, one applied edit sequence, one verification pass. Repeated across a manuscript, that is how an overwhelming rewrite becomes a book you can finish.