Sample Reports

Don’t take our word for it — read our work.

This battery was not built on theory. It was built on our own book — run over Maeve chapter by chapter through her editing, finding real corrections, which were made, and then run again, and again, until a full run came back clean. Only a machine proven in our own forge was ever offered to a client’s manuscript. The two reports below are that proof, kept: the first formal run, on the owner’s own novel, and the stress test that followed it, against a masterpiece.

A client’s report is never posted, excerpted, or sampled — not anyone’s, not ever. Clients receive their reports privately. The public samples on this page are only public-domain classics and the owner’s own books.

How We Test Our Own Machine

Before this battery ever touched a client’s book, we stress-tested it against a masterpiece: Mark Twain’s Adventures of Huckleberry Finn — 111,085 words, 444 pages, dialect on every page.

The battery found zero mechanical errors. It confirmed the craft — two deliberate dialect systems, never confused across 43 chapters — and it flagged two genuine seams for the author’s decision, both in Chapter XXI, where a lesser scan finds nothing at all. It also taught us four new checks, now permanent members of the battery. This calibration season alone taught the battery twenty-six new checks, and every one of them stayed. The battery is still growing — any genuine check a book teaches us joins it for good.

And it reported our own Gate in full, including the absolutes where this house and Mr. Twain part company. Nothing in the report is softened, and nothing in it is staged. The full run was one of ten: we stress-tested the battery on ten of the most difficult classics — over three million words (3,042,058 in all). We name only this one publicly; the rest were calibration, and calibration stays in the workshop.

Sample No. 1 — Adventures of Huckleberry Finn, Mark Twain

The Gate, Reported Plainly

Our integrity screening returned NOT CLEAR on this book — and the report says so on its first page, in plain terms. The Lord’s name passes without discussion: in 111,085 words there is no blasphemy, and the word “damn” appears nowhere in the book. Two of our standards the book does not meet: it is steeped in folk occultism as its characters’ working religion, and the period racial slur appears 208 times. On a live submission, that result ends the engagement at the Gate — deposit retained, manuscript returned untouched, this report as the explanation. Because the run was our own calibration, the full scan was completed anyway: a test that stops at the first failure teaches nothing.

What the Full Scan Found

• Two findings for the author’s decision — both in Chapter XXI. The rehearsal assigns the duke his Richard III sword fight and the king his Hamlet soliloquy; the printed playbill then swaps the assignments. And Colonel Sherburn swears he will endure Boggs “till one o’clock,” yet by the chapter’s own clock the shooting comes some minutes before the deadline he named. Our report lays out both readings of each seam and changes nothing: judgment calls belong to the author.
• Eight clusters of suspected errors — confirmed as deliberate craft. Two interlocking dialect systems held with machinelike steadiness across 43 chapters; the duke’s mangled soliloquy, a patchwork the text itself explains; money figures ($3,000, $6,000, a bag $415 light) that appear to quarrel and reconcile exactly; a handbill that contradicts the book’s own facts because the handbill is a forgery inside the story. A scan that cried error at any of these would be misreading the book.
• AI Sound-Alike reading: about one passage in fifty carries any of the surface patterns detection tools hunt — every one of them a set speech or a refrain with a human reason behind it. The report’s verdict, which we now print in every report we issue: an estimate is an editorial reading, never a verdict.

Service shown: Scan + Obvious Fixes, 444 pages × $2 = $888 total as-if value. Test run — no charge was made, and on a live submission the engagement would have ended at the Gate.

Report (PDF)

Read the Book Yourself, Free

Twain costs us nothing to share and serves every reader who lands here: the entire Adventures of Huckleberry Finn is free at Project Gutenberg — read it here. This shelf will also carry our own editor’s books, with their links, as each one publishes.

Sample No. 2 — Maeve, the Commanding Queen

The owner’s own novel.

The Twain run was the stress test. The first run was closer to home: our own editor’s novel, Maeve, the Commanding Queen — 73,607 words, 295 pages, the owner’s own work on the table.

The Gate cleared it. The battery fixed nothing on its own authority, and it brought the author two findings worth her decision: a character named some twenty-five paragraphs before his introduction, and a rhythm of chapter endings that had begun to lean on a single device. Three repeated lines were examined, confirmed as deliberate echoes, and left standing exactly as written.

• Gate: CLEAR. No use of the Lord’s name in vain, no blasphemy, no sacrilege; no devil or occult content; hard subjects handled in the open, without graphic detail, and the book lands on faith and redemption.
• Two items for the author’s decision — Daniel named in Chapter Eighteen, some pages before his introduction; and a signature chapter-closing device that clusters in the final four chapters. Both are reported with our reading of each; the author rules.
• Confirmed clean: every recurring name and Irish term spelled one way, every time; family ages, birth years, and intervals in agreement across all twenty-three chapters; every object the ending depends on planted early and accounted for at its return.
• AI Sound-Alike reading: about one passage in twenty carries any flagged surface pattern, clustered at chapter endings and in summary lines — the report names all five passages and calls them what they are: the author’s signature devices, flagged for awareness, not as defects.
• The price math, in the open: Gate Scan $50, credited in full on the CLEAR result; Scan + Obvious Fixes, 295 pages × $2 = $590 total value; as-if balance after the credit, $540. Test run — $0 due.

This is the actual report the run produced — the same document a Scan + Obvious Fixes client receives, on the owner’s own book, with nothing staged for this page:

Report (PDF)

We ran it on our own book first because that is the order honesty requires. Then we stress-tested it against a masterpiece.