Feature · Stage 3, Check the book

What your world gained and lost.

A revision that removes a subplot removes vocabulary with it, and it is easy to lose a term you meant to keep. Bring in the earlier draft, either the manuscript itself or a JSON export you saved, and the two term lists are set against each other.

Try it free
How it works

Two ways to bring in the earlier draft.

1

Upload the earlier manuscript

A .docx, .epub, .rtf or .txt of the older draft is read on the server for its list of term names and nothing else. No counts, definitions, contexts or exports come back from it, and the file is not stored. One earlier draft is read free with every paid scan.

2

Or load a JSON export you kept

The reader-editions screen exports the whole term list as JSON. That route is read in your browser rather than uploaded, costs nothing at all, and is still there once the free read for this scan has been used.

3

Added, renamed, gone, unchanged

Terms in only one of the two lists are listed as added or no longer present, with a count of the ones in both. Where a name that left is a near-match for one that arrived, the pair is pulled out as a probable rename rather than counted as both a loss and an addition.

Use cases

What it is for.

Structural edit

Confirming a cut subplot took all of its vocabulary with it.

Series

Seeing which of book one's terms survive into book two.

Co-writing

Checking what a collaborator's pass introduced.

Rename

Verifying a global rename actually reached every mention.

FAQ

Questions about comparing drafts.

Do I need to have exported the earlier draft in advance?

No. If you still have the earlier .docx, .epub, .rtf or .txt, upload it and it is read for its list of terms. A saved JSON export is the alternative, not the requirement.

Is the earlier draft uploaded?

It depends which route you take. An earlier manuscript is uploaded and read on the server, in memory, then discarded, and only the list of term names comes back from it. A JSON export is read in your browser and never uploaded.

How does it know something was renamed?

It does not know, it infers. Nothing is stored between scans, so there is no record of a rename anywhere, only two lists of names. A name missing from one draft is paired with the closest near-match that arrived in the other, by the same spelling-distance rule the similar-name checker uses, and the screen labels the pairs as probable rather than certain.

Does comparing use up one of my scans?

No. One earlier draft is read free with every paid scan, and a JSON export costs nothing at any time. Names alone come back from the free read, which is what keeps it from being a second scan by another name.

Why do I have to supply the earlier draft myself?

Because no manuscript or scan is ever stored on the server. There is no history to compare against, which is the same reason nothing of yours can leak from it.

See what changed since your last draft.

Free to preview, no sign-up required.