An editorial board for everything you write.
Rhēma reads a finished draft the way a demanding editor would — checks every load-bearing claim against real sources, maps the argument, and hands back a dossier you can act on.
No card. Your first review is included.
Start with a draft · Upload an outline · Start from scratch
What the board hands back
- Every finding quotes its passage
- Every verdict carries a URL that was fetched and read
- The second read never sees the first reader's notes
- Independent second read, not a panel of personas
A read-only sample dossier — the same worked example the Showcase ships.
Read-only sample. In your projects, each finding is something you can act on.
The Extinction of Man
The Editor's letter
This is a real 1897 essay arguing a real thesis — that humanity's tenure is not guaranteed — and it earns the reader's attention. The argument is honest; its evidence is not. The load-bearing natural-history claims are a century stale, and one is wrong by a factor of two on the axis the essay builds the horror on.
The piece also deserves credit where due: the central wager — dominant taxa fall without warning — survives contact with modern paleontology. The ledger below separates what survives from what does not, with the sources the review actually fetched.
Claim ledger
[9]Tier 1ContradictedContradicted · Tier 1
The evidence of the fossil Paradoxides shows creatures of this kind may attain a length of six feet.
Paradoxides' largest specimen is 37 cm (15 in); the largest complete trilobite ever found (Isotelus rex) is 72 cm — under 2.5 ft, not six. Paradoxides is also a trilobite, not a crustacean 'of this kind.' Wrong on the scale and the group.
[12]Tier 1Misleading in context
Man has swept away auk, bison, and dodo during the last two hundred years.
The bison was hunted to under 1,000 but never extinguished — today ~11,000 DOI-managed in 19 herds, designated the US national mammal in 2016; the dodo fell ~1681, outside the essay's own 200-year window; only the great auk (1844) fits.
[3]Tier 1Substantially accurate
In no case does the record of the fossils show a really dominant species succeeded by its own descendants.
The pattern holds in the mass-extinction literature — dominant clades are replaced by previously marginal competitors — though modern reviewers soften 'in no case' (post-extinction radiations do include survivor lineages).
[31]Tier 3Unverifiable
We may be unwittingly evolving some new and more terrible plague.
Speculative thesis claim — argued, not checked.
Findings
The six-foot crab is the essay's hinge — and it is wrong on the fossil record.
The nightmare scenario that sells the thesis is scaled off Paradoxides. The real genus tops out at 37 cm, and no complete trilobite has ever exceeded 72 cm. The essay's scariest object shrinks by half — the correction must be visible, because the fear is built on the number.
The essay's own timeline contradicts one of its three extinctions.
Within 'the last two hundred years' Wells lists auk, bison, and dodo as swept away. The bison survives (a conservation success, not an extinction) and the dodo predates the window. A second read, given only the draft, flags the same trio — the essay argues harder than its record.
The workflow, in lifecycle order
- Outline
The argument's skeleton, checked before prose.
- Draft
Your authorship — every suggestion waits for your approval.
- Review
The full board: claims, structure, discourse.
- Fact-check
Every checkable claim against fetched sources.
- Revise
Moves proposed as a response — you keep the pen.
Draft's suggestions never apply themselves — every one waits for your approval.
The author's seat
You read the dossier, pick the moves worth making, and watch the diff. Dispositions, apply-or-decline, a response letter when you revise — the instruments assume it's your book.
A register, not a person
A Voice is a set of registers and tics extracted from your own corpus — it tunes the review to your room, and a staleness pill tells you when the profile has drifted.
Runs you can watch
Every run carries its estimate-and-cap card live — the run never exceeds the cap you saw before it started, and the meter is on the page while it works.
Stay on the frontier
Powered by Claude — model tiers stay current as the frontier moves, and the rate card publishes what each tier costs before you spend it.
Show your work
The dossier's claims ledger is the proof: every verdict links the source it fetched, every contradicted claim shows the passage and the page. Judge the board by its citations.
Pricing
The same numbers the run sheet quotes — plans, credits, and the per-workflow cost cards.
Essay5,000 words · 20 claims — Review 2,600 – 11,730 credits
See pricing
Changelog
What's live now: the full editor, the public funnel, and the Showcase.
Questions
Does Rhēma write my essay?
No. The board reviews and proposes; the approval gates keep the pen with you.
Do I pay to look?
No card is required to start — your first review is included in the trial window.
Is my text used to train models?
Never. Drafts, letters, and telemetry stay out of training, and providers are bound by the same terms.