The moment every novelist dreads
It usually hits you in the shower, or two paragraphs into chapter 24. The name you picked nine months ago is wrong. Anna isn't sharp enough. Hannah is. You can feel the rest of the book settle around the new name within seconds — and then, almost immediately, you feel the dread of having to actually do it.
Because in most writing tools, renaming a character isn't a button. It's a project. You open Find & Replace, type "Anna" into the top box, type "Hannah" into the bottom box, hover over Replace All — and then you stop. Because you remember Annapolis. And Annabel. And the chapter where someone says "anna" as a common Italian word for grandma. And the dictated voice note from three weeks ago where you were figuring out her arc out loud, and it's saved as Anna, and now nothing will link it to her anymore.
So you don't click Replace All. You start clicking through every match manually. 47 matches across 23 chapters. It takes you the better part of an evening, you miss two, and you find them weeks later in proof copy.
Why traditional find-and-replace fails for novelists
The problem isn't that find-and-replace is bad. It's that find-and-replace doesn't know it's renaming a character. It treats your protagonist's name like any other string of letters, with no understanding of what she is to the book.
Three concrete failure modes:
- Substring matches. Replace "Anna" with "Hannah" and you get Hannahpolis, Hannahbel, Hannah Karenina, and an Italian grandmother named "hannah" mid-sentence. Scrivener's Project Replace has exactly this behaviour. There is a "whole word" toggle in Word, but it's off by default, and most novelists don't even know it exists.
- Case-insensitive collateral. "anna" is a real word in several languages. Lowercase the toggle and you start rewriting common nouns. Leave it on and you miss a few legitimate sentence-start instances. Either way, you're proofreading the whole book again.
- No idea what else needs updating. The character record in your bible, the voice-note transcripts that mention her, the audiobook synthesis prompts, the chapter wordcounts — none of these know you renamed anyone. They keep pointing at a name that no longer exists in the manuscript.
This isn't a UX nitpick. It's the reason most writers don't rename characters once they're 30,000 words in, even when they should. The friction wins.
How MimicReader handles it
The Writing Studio (Settings → enable Writing Studio) has a Characters tab inside every project. Each character you've introduced shows up as a card — name, role, brief description, and a Rename button.
Click Rename. Type the new name. Before anything changes, you get a dry-run preview:
If the count looks wrong — say, you expected 60 and you see 12 — you know something is off before you commit. Maybe half the chapters call her "Ann" instead, and you want to do two passes. Maybe you've got an alias in there. You cancel and check. Nothing has changed yet.
If the count looks right, you confirm. In the next second:
- The character record in the database updates from
name = "Anna"toname = "Hannah" - Every chapter's content is rewritten with the new name, preserving formatting
- Chapter wordcounts are recalculated (since the new name has a different character length)
- The old name is added to the character's alias list, so voice-note transcription still recognises it
30 seconds, end to end. No proofreading sweep. No missed instances. No Annapolis.
Step by step
- Open your project in the Writing Studio
- Click the Characters tab
- Find the character card, click Rename…
- Type the new name in the prompt
- Review the preview ("47 occurrences across 23 chapters")
- Confirm
That's it. The chapter list reflects the new name immediately. If you have the chapter view open in another tab, refresh — it's all updated.
The alias preservation matters more than you think
This is the bit that nothing else does, and it matters most for people writing with voice notes.
If you've been dictating notes for months — "OK, so Anna walks into the room and notices the…" — and you suddenly rename her to Hannah, every old transcript still says Anna. That's fine for searching old material. But when Whisper transcribes a new voice note, it has an initial_prompt that lists your characters by name. If that prompt no longer mentions Anna, then any new note you record that still calls her Anna (because you slipped, or because you were referring to her old name in context) might transcribe as "Hannah" or "Anya" or something else entirely. The model loses the anchor.
By keeping the old name as an alias, MimicReader feeds both names into Whisper's prompt: "characters in this story include Hannah (formerly Anna)…". The transcription still attributes Anna mentions to the right character. Your historical notes are intact. Your new dictations don't drift.
Honest limitations
Two things this doesn't do today:
- Case-sensitive by design. "Anna" and "anna" are different. The lowercase form is almost always a common noun in some language, and we'd rather miss the one legitimate sentence where lowercase Anna appears than rewrite hundreds of grandmas, names of saints, and Italian words for "year". If you genuinely want to rename a character who appears in lowercase too, do a second pass.
- Exact match only — Polish inflection (fleksja) is not handled yet. In Polish, a character named Ezra appears as Ezry, Ezrze, Ezrą depending on grammatical case. The current rename catches "Ezra" but not the inflected forms. We know. Inflection-aware rename is on the roadmap and we'll ship it once we can do it without false positives across 23 languages. For now, Polish-language manuscripts need a manual sweep for the inflected variants after the main rename.
Both are deliberate choices to ship a tool that's safe by default rather than one that's clever-but-occasionally-destructive.
Compared to the alternatives
| Tool | Character-aware? | Preview before commit? | Whole-word default? | Updates aliases? |
|---|---|---|---|---|
| Word / Google Docs | No | No (one match at a time) | No (toggle exists, off by default) | No |
| Scrivener (Project Replace) | No | No | Configurable | No |
| Atticus | No | No | No | No |
| MimicReader Writing Studio | Yes | Yes (count by chapter) | Yes | Yes |
Scrivener's Project Replace is the closest competitor, and it's still a generic find-and-replace dressed up with a project-wide scope. It doesn't know about your characters because it has no concept of characters. The Writing Studio does — your characters are first-class records, with names, aliases, descriptions, and a button per card.
One small button, a lot less grief
A writer should be able to change a character's name without it becoming an evening's work. They should be able to see what will change before it changes. They should know that their voice-note transcripts won't suddenly stop recognising the protagonist. They should be able to do all of that and then get back to the chapter they were writing.
This is that button. It's small. It's been requested by literally every novelist who's tried the Writing Studio. It took us less time to build than to write the blog post explaining it. And it's there now, on every character card in every project.
Try the Writing Studio free
Settings → enable Writing Studio. Characters tab in every project. Rename, preview, confirm — 30 seconds. Included in every free account.
Start Free