Protect your search traffic before you publish.

Anyone can publish a thousand pages a day. No one can see which ones get the site demoted.

• AI is now the default writer of the web — volume is infinite, quality goes unverified.
• Google scores quality at the domain level: a handful of weak pages can demote every page you rank.
• Even brands with human writers are exposed — 75%+ use freelancers, and manual review can't read every page.
• A demotion erases 30–80% of organic traffic for 6–24 months — and today everyone ships blind to it.

Content went infinite. Google started policing quality site-wide.

The exact moment a market needs a gate is when publishing becomes free and the penalty for bad output becomes severe. Both just happened.

A quality gate every page passes before it goes live.

ArticleGrade grades a page 0–100 across E-E-A-T, AI-artifacts, factual depth and SEO structure, and returns a pass/fail verdict with phrase-level fixes — as an API call, in the dashboard, on a pasted draft, or right inside the CMS.

Not a demo — a launched product with five ways in.

The same engine ships through five surfaces, so the gate meets teams wherever they publish. Everything below is live in production today.

The moat is the system, not the prompt: a live-fetch harness, a 26-gate rubric, a dual-LLM cross-check, and a remediate-and-re-audit loop — none of which a single model call reproduces.

Battle-tested internally, now launched and monetizing.

We built ArticleGrade as the pre-publish gate inside our own AI-content pipeline and ran it on real traffic — before a line of sales. It is now a public product with live billing.

What launch de-risked: the product works and bills. The round is now about proving the funnel — the free grader is metered and instrumented, so free-to-paid conversion becomes a measured number, not a benchmark.

A wedge into a budget that already exists.

A beachhead story, not a share-of-TAM story: we win one acute segment first, then expand to every business that publishes. Bars illustrative; the % is to scale.

A crowded edge — and an empty center.

Land self-serve, expand to volume.

A free grader funnels into five paid tiers — priced so the entry point is a rounding error and volume carries the ARR. All live in Stripe today.

Overage is a flat $0.45 / deep audit past a plan's allowance. Two-stage screening — a free first pass on every page, a deep audit only where it counts — holds gross margin near 80% today, with a path past 90% as inference gets cheaper.

Honest margins. Payback inside a year.

The honest math: screens are free to the customer and cost ~$0.002 to run; the deep audit costs ~$0.14 against the $0.45 overage price — ~69% — and is the one real variable cost, and the single lever past 90% as inference cheapens. Hosting is a fixed ~$75/mo VPS, not a per-audit cost. During the Sonnet-5 intro window the deep cost runs ~$0.09, a temporary margin tailwind.

Compounding to a ~$2.9M ARR target by Year 3.

How to read it: the $2.86M bull is the ceiling this compounds to if the agency channel and free-grader funnel fire as designed. The $1.28M base is what this $150K is underwritten to — ~$355K ARR / ~150 active accounts by month 18. Bear $0.40M is the downside. Every figure is built bottom-up — accounts × ARPU × retention — not a top-down slice of TAM. The bull is the target; the base is the promise.

12% of the logos carry ~68% of the revenue.

Base case, by quarter. A self-serve long tail builds logo count and proof, but ~60 Scale and agency accounts by Year 3 drive roughly two-thirds of ARR — the mix that makes a ~$2.6K blended ARPU hold up. The new $29 Starter tier widens the top of the funnel further.

The three-year base case.

The Year-3 base sits between Originality.ai at ~$2.3M and the slow Clearscope case — a defensible middle, not a hockey stick. Bear $0.40M · Bull $2.86M — the spread is funnel conversion, not market size.

Win the operators who feel the pain first — then widen.

• Wedge — an API pre-publish gate for AI-content-at-scale operators and their agencies, sold as demotion insurance, not a writing tool.
• Motion — metered free grader → self-serve ($29 / $79) → API & agency white-label → brand content teams. Each tier widens the buyer; the extension and WP plugin are zero-friction distribution.
• Comparable — Surfer reached ~$15M and Clearscope ~$1.5M on under $1M of marketing — this category buys on ROI, not ad spend.
• Window — roughly 12–18 months before SEO optimizers bolt live-fetch QA onto their suites. We own the gate first.

Beyond AI operators: cheap insurance for any content team.

Every business that publishes text shares one exposure — a single weak batch can demote the whole domain. The same gate that guards AI operators guards any content team, for a fraction of the editor it replaces.

Why it's now baseline risk: 52% of new content is AI-written, 90% of marketers ship AI and ~20% goes unreviewed — against organic that drives ~40%+ of revenue, with near-zero recoveries in a 50-site study. Documented site-wide demotions: HouseFresh −91%, DMARGE 8M→300k, Retro Dodo −85%. The editorial-QA budget is already real — $3–8B, anchored by Grammarly ($13B), Writer ($1.9B), Acrolinx. (Honest: these losses are proportional, not guaranteed — we lower the share of weak pages.)

Product is shipped. $150K de-risks the funnel.

Because the product already ships and bills, we're funding a go-to-market experiment with three falsifiable proof points — live funnel conversion, signed logos, and a measured CAC — not asking you to take product risk on faith.

The product is built. 85% of the raise buys distribution.

Only 10% ($15K) goes to product & engineering — the gate ships and bills today, so the raise isn't building it, it's selling it. 85% ($127.5K) goes to growth across five channels; 5% is an operating buffer.

Operating buffer 5% ($7.5K). Product-light, growth-heavy — because the risk left to retire is go-to-market, not the product.

$127.5K of growth → ~197 new logos, at a 3.2× return.

Bottom-up, by channel — budget ÷ CAC = logos, logos × per-channel ARPU = ARR — grounded in a measured ~$647 blended CAC, held flat, not heroically compressed to flatter the return.

How it reconciles — and where the risk sits: this is the new-logo acquisition engine, not the full P&L. $127.5K buys ~197 gross logos — ~150 net active by month 18 after self-serve churn — and ~$200K of new-logo ARR at a blended $647 CAC. Content is the largest, most durable line (~35% of channel ARR across 87 owned, dogfooded logos); partnerships are the highest-ACV bet (~$3,800 ACV, ~27% of ARR from 14 logos) — assumed, not yet contracted, and the one number most worth pressure-testing. The bridge from ~$200K of new-logo ARR (150 accounts valued at first-year ARPU) to the ~$355K base-case 18-month ARR is account maturation — the same accounts expanding as seats and metered usage grow — the biggest upside and the biggest thing still to prove. Every line is throttleable: partnerships and the funnel run as staged probes with kill/scale gates on the first ~10 signed agency logos and real funnel-conversion data before more capital goes in.

Underwriting: the $2.86M bull is the Year-3 ceiling this engine builds toward; the round is underwritten to the $1.28M base. Numbers are projections built on measured unit economics — not guarantees.