Technology

Foolscap Says It Keeps Track of Your Whole Book

Martin HollowayPublished 6m ago4 min readBased on 1 source
Reading level
Foolscap Says It Keeps Track of Your Whole Book
source:foolscap.works

Foolscap, a writing app, says it keeps the entire manuscript in view while an author drafts. Its website, published on 19 November 2026, describes a tool that treats the whole book as one working document, rather than handling each chapter or scene as an isolated file. Foolscap

The site also says the app notices what connects and what is left hanging as the user writes. Those two functions form the core of the pitch. A thread introduced early, a motive that never resolves, or a reference that points to nothing would, according to the product's description, be flagged during drafting rather than found in a later revision pass. Foolscap

The page is explicit on one boundary: Foolscap never writes a word for the user. Its role, as described, is to observe and flag, not to generate prose. Foolscap That distinction is the central design claim, and it is the one most likely to shape how working authors compare the product with the many writing tools built around text generation.

The product targets book-length work, since the site frames its capability around the whole book rather than a single document or passage. That scope differs in kind. Spellcheckers and grammar linters have long operated at the level of the sentence. Continuity tracking across hundreds of pages has mostly fallen to professional editors, to writers' own notes, or to memory. A tool that maintains a running model of a full manuscript and reports on its loose ends is tackling a different class of problem.

The broader context here is technical as much as editorial. Holding a book-length text in working state and checking cross-references across it is a question of retrieval, indexing and consistency checking, not only of language modelling. A context window, the amount of text a model can consider at once, is one option. An index built over the draft is another. Each approach carries different trade-offs in latency, cost and the reliability of the flags it raises, and the site does not say which one Foolscap uses.

In my view, the 'never writes a word' constraint deserves as much attention as the feature list. Many writers already use general-purpose assistants, and the common worry is that the machine's sentences gradually replace the author's. A tool built to point at problems and leave the repair to the writer keeps the question of authorship where it started. Whether that holds up over a long project and under deadline pressure is for users to test.

The case in its favour is straightforward. A writer gets a second reader that does not tire of checking chapter forty against chapter one, and that catches an unpaid setup before a reader does. The cautionary case is equally clear. A flagging tool is only as useful as its precision, and false positives on a manuscript with deliberate ambiguity could distract as much as the gaps they were meant to catch.

What the material establishes is modest: a product exists that presents itself as whole-book aware, identifies connections and open threads during drafting, and declines to author text. How well it performs, who is using it, and how it is built will depend on what it does once writers have it in front of them.