Cited · Three modules
Your expertise is already written down: in Word files, PDFs, directory profiles and partners’ heads. I turn it into a fact-checked knowledge hub on your own domain, keep it growing with one-click approvals, and send it out as a week of on-brand posts per channel.
Proof on file
CLIENT · NAME CHANGED1,100+
pages live on the firm's own domain
Corvane Legal, a German law firm: articles captured, reworked, claim-checked, approved and published on the firm’s own domain, then handed over into its own Cloudflare account.
observed 2026-10-05 · live sitemap count
03 Fit
Built for you if
Not for you if
04 The problem
The law firm behind my flagship case had thousands of expert articles on a third-party lawyer directory. Every one of them built the directory’s authority, not the firm’s. When someone asked an assistant about the firm’s speciality, the answer didn’t mention the firm.
Most expert firms have a version of this: a Word file of answers nobody published, a PDF library behind a login, a partner who explains the same thing on every call. The bottleneck isn’t knowledge. It’s turning that knowledge into pages and posts every week, safely, without hiring a content team.
And plain AI writing makes it worse. On one batch, the AI suggested 1,175 internal links; 72 were hallucinated, with anchors that weren’t in the text and pages that didn’t exist. That’s why nothing in this engine publishes without passing a gate.
Signs you need this
06 Three modules
Each module stands on its own and plugs into the others. You never pay twice for the same part.
01 · Capture
Your back catalogue published as a searchable knowledge hub on your own domain, right beside your current site.
02 · Publish
Continuous capture, rework, claim-by-claim fact check, one-click approval and publishing, running on its own every few minutes.
03 · Distribute
A week of on-brand posts per channel, drafted from your published expertise and approved from a phone in minutes.
Plus running costs of ~€30–150/mo, paid directly to the AI, scraping and hosting providers.
07 What you get
The full engine includes everything below; Capture and Distribute can be bought on their own and joined up later.
Your back catalogue as fast, searchable pages with sitemap and schema, served inside your live header and footer.
Every factual statement in a reworked article checked against its source as supported, partial or unsupported. Unsupported claims are flagged and held for a person, not softened.
AI-suggested links kept only if the anchor exists verbatim, the target page exists, and a relevance check passes.
A daily email of what’s ready, with signed single-use approval links and shareable preview links.
A hard monthly cap on AI spend and a daily rework limit. When the cap is hit, the engine stops; it doesn’t send you a surprise bill.
Which AI crawlers read which pages, by name, with an alert if they stop coming.
Every enquiry tagged with the article that brought it in, carried in the link without cookies.
A week of on-brand posts per channel, approved, declined or commented on from a phone. Feedback shapes the next draft.
Automated tests on every pipeline stage, a runbook, and the whole system moved into your Cloudflare account.
08 How it runs
It starts in pilot mode: you approve every article from the daily digest. Automatic publishing is switched on only for the categories you choose, and you can switch it off again at any time.
Agents at work Human gate: nothing moves on until a person approves
Step 1: Week 1
Every source inventoried: documents, archives, directory profiles. Tone, banned phrases and things that may never be claimed are written down.
Human gate You approve the source list and the editorial rules.
Step 2: Weeks 2–3
Sources archived and parsed, then imported. The hub goes live beside your site on one path, with your header and footer.
Human gate You approve the hub before it’s public.
Step 3: Weeks 3–5
Articles rewritten into answer-first pages; every claim checked against the source. Anything unsupported is flagged and held for a person, not softened.
Human gate Pilot mode: you approve every article from the daily digest.
Step 4: Weeks 5–7
Per-bot crawler analytics, article-to-form attribution, the budget breaker and its alerts.
Step 5: Weeks 7–9
Moved into your Cloudflare account with CI deploys, runbook written, team trained on the digest.
Human gate You decide when, and for which categories, approval becomes automatic.
09 Proof
The law-firm engine is a client system, live and handed over. Cuckoo is my own distribution engine, in pilot while the social platforms approve its publishing access.
1,100+ pages live
Plus 230+ on a sister domain. 649 automated tests, a 5 min heartbeat and 5 domains on one engine.
Core built in ~2 days
Drafts a week of on-brand posts per channel with 4 AI providers; clients approve from their phone. Publishing integrations are built and waiting on platform approvals.
10 Price
Each module has its own published starting price. I usually suggest starting with Capture, proving the hub works, then adding Publish or Distribute.
Costs more when
Costs less when
Paid by you, directly to each provider, in your own accounts. I add no markup and no hosting fee.
| What | How it works | Paid to |
|---|---|---|
| AI model usage | Rework, fact checks and drafts, billed by the AI provider. Throttled by a daily limit and capped by your breaker; the law-firm engine runs at ~€33/mo. | Anthropic / OpenAI |
| Scraping | Only if your sources sit on third-party sites. Normal operation usually fits a provider’s free tier. | Scraping provider |
| Hosting, database, cron | Cloudflare Workers, D1 and KV. Small volumes fit Cloudflare’s free plan; Workers Paid is budgeted when volume or CPU needs it. | Cloudflare |
| Digest emails | Cloudflare Email or the provider you already use. | Cloudflare / yours |
Typical running costs: ~€30–150/mo, paid directly to the providers. The breaker defaults to a hard €100 a month.
11 Handover
No licence fees to me, no lock-in, no proprietary platform. If I disappeared tomorrow, any competent developer could pick it up from the runbook.
12 Questions
Rather talk it through first?
20 minutes with me, not a sales rep. You leave with a clear next step — whether or not we work together.
Text extracted from the HTML this server sends — no JavaScript executed, the way GPTBot, ClaudeBot and PerplexityBot fetch pages. It is also the Markdown twin I publish for AI agents.
Loading…
<!doctype html> <title>Custom Web Tools & Cookie-Free Websites | AIWebRevolution</title> … <body> <div id="root"></div> <script type="module" src="/assets/index.js"></script> </body> → 0 words of page content. Beautiful to people. Blank to ChatGPT.
Two minutes · personal reply