Most AI companies ask for your trust. We built a house that can show you its building inspection instead. This page is that inspection — written plainly, with nothing on it our own gates did not verify.
The three inspections every room must pass
1 · The Building Code
Every module of every system is scored against a written engineering code — architecture, evidence discipline, testing, security, observability. The bar applies to everything; three principles are foundational, and a single zero on any of them fails the whole build.
2 · The Promise Audit
Every claim on this website is mapped to the machinery that produces it. A promise without machinery gets exactly two options: build it, or take the promise down. Nothing stays claimed-but-undelivered.
3 · The Elevation Pass
Design held to frontier standard: structured hand-offs, self-verification before delivery, quality measured as a number, honest behavior when something is unavailable, and a system that records its own outcomes so it improves instead of decays.
What stands today — verifiable, not claimed
One shared kernel, eight rooms. The constitutional modules (owner control, verification, the record of every change) exist as one canonical version across all eight systems, pinned by hash in a manifest and re-checked on every verification run. Drift is detected by machine, not memory.
One command verifies the whole house. A house verifier re-runs every module's self-test and every room's diagnostic — 156 modules as of this writing, all passing, and the 3 that depend on optional local software declare an honest hold rather than pretending. It writes a dated ledger every time.
All eight rooms — certified, sealed, and stamped. Every room now runs a conformance gate scoring every module against the code, an append-only flight recorder fingerprinting every run, and a red seal that refuses to ship if any gate, test, or safeguard fails — proven by planting violations and watching each room refuse. On July 12, 2026 the full house verified 156 of 156 modules passing — zero holds — and the owner's stamp, issued through a gate that bars self-certification and demands a distinct verifier of record, returned FULLY VERIFIED across the estate.
The first screenplay exists. A complete novel converted to a feature screenplay by the house's own machinery: 101 rendered pages — inside the hard 92–120 band, verified by the room's own gate — after six professional revision passes (producer, cinematographer, network executive, award screenwriter, cadence, fidelity), every beat traced to its source passage. It is the founding reference of the room's golden set.
Independent skeptical review — with the verdicts kept. Nothing certifies itself here. Fresh reviewers who did not build the work grade it against the code, and every review attempt is logged — including the first one, which was a refusal. We keep our failures on the books, because a review log with only passes in it isn't a review log.
Quality as a number, not a feeling — in every room that scores. Golden sets of reference readings and consistency instruments guard every verdict and conversion room: they measure whether repeated readings of the same work agree, and the instrument refuses to publish any number it has not yet earned. As the measurements accrue, they will be published here. No one else in this market dares to measure this. We built the instrument first — then we put it in every room that delivers a score.
The first live verdict is delivered. A full-length manuscript, read whole — every finding anchored to its location with the author's own words quoted, comparable titles verified against live sources (and the ones not machine-checked, flagged as such — in the report itself). The receipt exists.
The Story Seal is live in both companion rooms. In Writing Your Story and Writing My Words, a sealed story opens only for its teller. No summaries to staff, no reports to family, no exceptions except those the law itself requires — the seal is machinery, not a policy sentence.
The Pen is live in both companion rooms. For the teller who would rather talk than type, The Pen writes it for them — truly their words, never ours — elevated, read back, and placed on the page exactly as they want it there.
The ten-revision license law is wired in code in both screen rooms. In The Screenwriter's Room (manuscript to film), the embedded license code carries the findings brief, the walkthrough, and edits on the spot. In Showrunner’s Room (manuscript to television), the same law holds, episode by episode. Up to ten revision sessions, counted by each room’s own license gate. The eleventh session refuses, plainly, with the completion message. Stated before the work begins, enforced by the machinery, not the fine print.
The Honest Read™ — the read without a bad day in it
Every human read arrives with a human attached — a stressful week, a genre they quietly hate, a character who reminds them of someone. None of it is malice. All of it lands on your pages anyway.
Our rooms rate what is on the pages, measured against industry standard with no human emotion attached — the thriller and the faith novel, the love story and the war story, judged by the same steady standard. Not because of a bad day, stress, or a bad break — your work gets the score it earned. Every time.
And because a score alone is not a service, every finding arrives taught — the why, the where, the how, framed as “you may wish to consider” — and the authority stays yours: ACCEPT · ADAPT · DECLINE. We sit with you. We never write over you.
Every claim on this site has been researched and verified. Our comparisons come from live market research. Our formats come from produced, award-winning scripts. Our reads are judged against the industry standard — never against a reader's taste. Our suggestions are professional opinions grounded in the industry. Our walkthroughs are by your choice only. Your voice arrives yours and leaves yours. A service for the writer first. That will not change.
The platform this was proven on
Built, verified, and proven on Claude (Anthropic) — and only there. Every gate, every test, every ledger, and every receipt on this page was earned on Claude. We recommend it as the platform for every room because it is the only platform we can stand behind: we have not run this house anywhere else, so we make no promise about how it behaves anywhere else. The promises on this site apply as delivered, on the platform they were proven on.
What we will not do
- Publish a claim our machinery cannot demonstrate. (The Promise Audit exists to catch exactly this.)
- Let the builder grade its own work. (Fresh-context review, verdicts verbatim, no re-rolls.)
- Use your manuscript to train AI models. (Commercial API terms; your pages remain yours.)
- Quietly weaken a gate. (Every change lands in an append-only record; sealed packages are byte-verified.)
Asked plainly, answered plainly — the questions serious buyers ask
When does the Greenhouse open?
The house opens to the public in Q4 2026. Beta testing across all eight systems begins Q3 2026 — real writers, real manuscripts, the rooms as they will ship. A beta seat is requested, not bought: write to info.lovingmywords@gmail.com and tell us what you’re writing.
Can I use the rooms in my own language?
Yes. Reach out in the language you write in, and the rooms answer there — and your own words, your manuscript, your telling, stay in your language, always. The house is strongest in the major world languages today, and we widen that with every writer who joins us. Two things stay anchored on purpose: the sacred promises and your legal terms keep their filed wording until a careful human translation is added — because a machine should never quietly reword a promise. The voice you arrive with is the voice you leave with, in the language you think in.
Will the AI write my book for me?
No — and if that’s what you want, we’re honestly the wrong house. We don’t write your story for you; we help you discover what it’s capable of. Every suggestion is yours to accept, modify, or decline, and the voice you arrive with is the voice you leave with. Always.
Then what am I actually paying for?
A professional development process that used to be expensive, gatekept, and opaque: beta reads that quote your actual pages, editing where every decision is ruled on by you, and conversions that turn a finished book into an industry-formatted screenplay or television season — with inspection reports you can read on this page.
How long does a real manuscript edit take?
Plan on 5–10 hours a week of your own time, for roughly 2–3 months, with real edits and real revisions. This is masterclass-level work on your book, not a weekend car wash. Some manuscripts move faster; honest work takes what it takes, and we will never pretend otherwise.
Isn’t this just ChatGPT with extra steps?
Test that claim: generic tools give generic notes and verifiably do not quote your pages. Our rooms read the whole work, anchor every finding to your own words, refuse to ship their own failures, and carry license, privacy, and honesty machinery you can inspect above. See the comparison — then judge for yourself.
Who owns what I write? Will my manuscript be used to train an AI?
You own everything — the manuscript you brought and every page you leave with. Your work is never used to train anything, never shared, never republished. It’s in our terms in plain English, in writing.
Can I sign up and pay right on this site?
Not yet — by design. The house is in its beta season: rooms open door by door, through us, so every early writer gets a certified room and a person watching how it serves them. Write to info.lovingmywords@gmail.com and tell us what you’re working on. Public signup comes when our legal review completes — and not one day before.
What do I need to use it?
A computer or a phone, and a Claude account (the AI platform our systems are certified on — see “The platform” above). We provide the room itself, set up and verified. No installs, no technical skill.
Can I talk instead of type?
Yes. Every room accepts the microphone button in the Claude app, and the companion rooms carry THE SPOKEN PEN — one big button, you talk, your words appear in large print, and the room writes with you. Built for hands that shake and eyes that tire. Works on computers and phones.
Are my children safe using this?
Our rooms are built for adults — writers, veterans, and elders — and the platform they run on requires users to be 18 or older, so we’ll say it straight: this isn’t a children’s product. The rooms never introduce content beyond the manuscript the author brings, and the companion rooms carry distress-aware safeguards. A family capturing a grandparent’s story does it the way it should be done anyway: together, with an adult at the table.
What if the verdict on my manuscript is harsh?
It will be honest, which is different from harsh — and it will quote your own pages so you can see exactly why. That’s The Honest Read™: the read without a bad day in it, delivered with respect, ruled on by you. We sit with you. We never write over you.
What stops someone from running five different books through one purchase?
Machinery, not trust: at intake your work receives a license code and a structural fingerprint. Rename your book freely — revisions of the same story pass all the way to the finish. A different story under an old code is flagged and refused with the reason stated. One purchase, one work, start to finish. A new book is a new license.
What happens when my revisions run out?
Each license includes ten full revision sessions — masterclass depth. At the cap, the room says so plainly and your finished deliverables are yours to keep, forever. Nothing expires quietly, and nothing you made is ever held hostage.
For television: what does “per-episode” mean?
Your series license covers the series design and episode one. Each following episode is added on its own, built one at a time, in order, with its own revision sessions — so you pay as the season grows, never for episodes that don’t exist yet.
I’d be writing on a shared computer. Is my story private?
Yes — the Story Seal locks each teller’s story under their own code, so on a facility’s common-room computer your pages open for you and no one else. No summaries to staff, no reports to family. Their story stays theirs.
Do veterans really never pay? What’s the catch?
They never pay. Not a trial, not a discount, not a first-month-free: if you served, you will never pay to tell your story here. The house stands behind that. There is no catch — it’s the reason this place was built.
How accurate is the AI? Will it make mistakes?
“No AI is 100% accurate throughout — our goal is to be the first.” That sentence runs this whole page: every promise names its machinery, every mechanism proves itself with tests, and when something can’t run, the system says so instead of pretending. When we can’t verify a thing, you’ll hear that from us first.
No deliverable is complete until you believe it’s ready.
Every manuscript, screenplay, series, report, or review we create is designed to bring your story as close to your vision as possible.
Because our systems are built with artificial intelligence, they can occasionally miss a detail or interpret something differently than you intended. Just like working with any editor, collaborator, or creative partner, refinement is sometimes part of the process.
That’s why every deliverable includes the opportunity to review the work, make changes where you feel they’re needed, and walk through those revisions with us.
We don’t believe perfection comes from one pass.
We believe it comes from working together until the story feels right to you.
Our goal isn’t to claim 100% perfection.
Our goal is to build systems that continually move closer to it — while always protecting your story, your voice, and your final decision.
For diligence
Partners and institutions may request the underlying records under NDA: verification ledgers, review logs, conformance scorecards, and the sourced market research behind every figure we publish.
Request the records