GPT-6 Astra (Max) Build4 pages One complete homepage + three...

Prompt

GPT-6 Astra (Max) Build4 pages One complete homepage + three standalone policy pages.Put contact information on the homepage and in the footer—no separate Contact page needed. Page Must-have sections/components Home/ Intro:who you are, who you help, what you sell.Offers + pricing:deliverables, exact prices, currency, one-time/recurring billing, delivery/access timing.Proof:genuine testimonials, contextualised results, authorised logos.About:short founder bio and business identity.Action/contact:one primary buy/book CTA and a public business email. Privacy/privacy Data collected, purposes, cookies, processors, retention, rights, contact details. Describe your actual practices—not generic boilerplate. Terms/terms Legal business name, purchase/use/licensing terms, payment and delivery, renewal/cancellation where relevant, liability and governing law. Refunds/refunds Eligibility, refund window, exclusions, request process, contact and processing times. Align with processor terms and applicable consumer rights. Structural and technical choices • One-page front door, separate legal URLs.Use stable anchors such as/#pricing,/#proofand/#contact. Link all policies from every footer and near purchase CTAs. • Fast, mobile-first HTML.Static or server-rendered pages, HTTPS, compressed images, minimal JavaScript. Don’t hide essential information in screenshots, popups or login-only areas. • Extraction-friendly content.Descriptive headings, named offers, explicit prices and short factual paragraphs. Keep business names, offer details and prices consistent across the site and checkout. • Basic discoverability.Crawlable links, descriptive page titles, canonical URLs, sitemap, and no accidental search/retrieval crawler blocks. • Useful JSON-LD only.PersonorOrganization, plus appropriateProduct/ServiceandOffermarkup. Match visible content. Schema helps machines interpret facts; it doesnotguarantee AI citations. Skip:blog, separate About/Pricing/Testimonials pages, elaborate animations and a heavy CMS. Processor caveat:Confirm current product eligibility first—standalone consulting/agency services may not qualify for Paddle or Lemon Squeezy. Complete website pages alone do not guarantee approval. Claude Opus 5.5 (Max) 5 pages 1. Home/ 2. Pricing/pricing 3. Terms/terms 4. Privacy/privacy 5. Refund Policy/refund-policy Add a sixth (Impressum) only if you're based in Germany or Austria. Skip the blog and any separate About, Contact, or service pages. Check this before building:Paddle and Lemon Squeezy are built for software and digital products. Both restrict human-delivered services (consulting, coaching, done-for-you). If that's what you sell, confirm eligibility in their acceptable-use policies first. If services aren't allowed, sell them through Stripe and use Paddle or Lemon Squeezy only for digital products. Structure: hybrid Home is a single scroll for prospects. Pricing and policies get their own URLs because reviewers look for them, checkout links to them, and AI engines cite specific pages ("What does [you] charge?" →/pricing). Must-haves per page Home • Hero:H1 states what you do, for whom, and the outcome. Directly below: "[Name] is a [role] who helps [audience] get [result]." Use one CTA and repeat it after the proof strip and at the bottom. • Proof strip:client logos (with permission) or 2–3 hard numbers. • Offers:2–3 cards with name, one-line outcome, price, and a link to/pricing. • Results:3–5 testimonials (full name, title, company, photo) plus 1–3 mini case results with numbers and timeframes. • How it works:3 steps. • About:real photo, 3–4 line bio, credentials, profile links. • FAQ:5–8 real buyer questions. • Contact:email in plain text, not just a form. • Footer (all pages):legal name, email, links to Pricing/Terms/Privacy/Refunds, and "Payments processed by [Paddle/Lemon Squeezy], our Merchant of Record." Pricing • H1 "[Brand] Pricing" plus a "Last updated" date. • For each offer: ◦ name ◦ exact price and currency ◦ one-time or recurring (with billing interval) ◦ what's included and who it's for ◦ delivery timeline ◦ Buy button linking to checkout • Show real prices. No "book a call for pricing." • An HTML comparison table if you have 2+ offers. • A refund summary linking to the full policy. • 1–2 testimonials near the Buy buttons, plus a short FAQ (billing, cancellation, invoices, tax). Terms • Legal name (yours or your entity, matching your Paddle/Lemon Squeezy account), contact, effective date. • What you sell and how it's delivered; payment, renewals, cancellation; license/usage; liability; governing law. • Merchant of Record clause. Paddle's standard wording:"Our order process is conducted by our online reseller Paddle.com. Paddle.com is the Merchant of Record for all our orders. Paddle provides all customer service inquiries and handles returns."Use the Lemon Squeezy equivalent if that's your processor. Privacy • Who you are, what you collect and why, processors (MoR, analytics, email, scheduling), retention, user rights (GDPR/CCPA), cookies, contact, effective date. Refund Policy • Exact window in days, eligibility, how to request, who processes it (the MoR), timeline, subscription cancellation, effective date. • Don't write "all sales final." Paddle expects a real refund window; 14 days is the safe default. Technical choices • Static HTML:use Astro, or Framer/Webflow if you want no-code. Most AI crawlers don't run JavaScript, so content must be in the raw HTML. Avoid client-rendered single-page apps and hash-URL "pages" (e.g., Carrd sections). • Lean and mobile-first: ◦ single column, one font, one accent color ◦ compressed WebP images, no animation libraries ◦ link to Calendly/Cal.com instead of embedding ◦ cookieless analytics (Plausible/Fathom), which usually means no cookie banner ◦ target 95+ on Lighthouse mobile • Extractable writing: ◦ one H1 per page and one H2 per section ◦ open each section with a plain factual sentence ◦ give each page a unique title and meta description ◦ put prices, names, and claims in text, never in images • Consistency:use identical offer names, prices, and one-line bio across your site, schema, checkout, and LinkedIn/X. Reviewers compare the site to checkout, and AI engines trust facts that repeat. • Schema (JSON-LD): ◦ Home:Person(plusOrganizationif you use a brand name) withname,url,image,description,email, andsameAs(your profiles). AddWebSite. ◦ Pricing:Service(orProductfor digital goods) withOffer(price,priceCurrency,url). ◦ FAQPageis optional. Google no longer shows FAQ rich results for most sites; AI pulls from the visible Q&A. ◦ SkipReview/AggregateRatingon your own testimonials. Google ignores self-serving reviews. • Crawl access:a permissive robots.txt and a sitemap. If you use Cloudflare, check its AI-bot blocking setting. Submit to Google Search Console and Bing Webmaster Tools (ChatGPT search leans on Bing). • Credibility basics:custom domain, HTTPS, email at your domain, favicon, and an Open Graph image/title/description. The link preview in a prospect's DMs is your first impression. • Optional:/llms.txt. It's unproven, but takes 10 minutes. Before you submit for review • No "coming soon" sections and no dead links. • Legal links in the footer. • Names and prices on the site match your Paddle/Lemon Squeezy products exactly. • Re-check both platforms' current checklists, since they change. Claude Fable 5.1 (Max) 1. Page count: 5 (+1 per offer if you sell more than one) Home — the one-pager that does the selling About Terms Privacy Refund Policy 6+. /offers/[name] — only if you have 2+ distinct offers. One offer = Home is the offer page. More than 3 offers = consolidate. Realistic total: 5–8. Don't build: a blog, a contact page (section + footer email is enough), a separate pricing page, a case‑study library, a thank‑you page (the MoR sends receipts). 2. Must‑have sections Home (in this order) Hero — one sentence: who you help + outcome. One sub‑line: how. Primary CTA (Buy/Book), secondary (See pricing). Trust bar directly beneath: logos or 2–3 hard numbers. Offers + pricing — one card per offer: name, who it's for, exactly what's included, format/delivery/timeline, price with currency and billing type (one‑time/monthly), CTA opening the Paddle/LS overlay checkout. This single section clears the "product info + pricing" check for both MoRs. Proof — 3–6 testimonials with full name, role, company, photo/logo, and a specific result. 2–3 one‑line results (before → after, with numbers). How it works — 3 steps: pay → onboard → deliver, with timing. About (short) — 2–3 lines, photo, link to /about. FAQ — 6–8 questions: refund terms, delivery time, who it's not for, how billing works ("processed by Paddle/Lemon Squeezy as Merchant of Record"), support channel. Final CTA Footer (every page) — legal business name, email on your own domain, country/address, Terms · Privacy · Refunds links, socials. About H1 "About [Name]" + photo First line in entity form: "[Name] is a [role] who helps [audience] [outcome]." Reuse this exact sentence in the hero, meta description, and social bios. Track record: years, numbers, notable clients, credentials What you do and don't do Links to LinkedIn/X/YouTube Email + CTA Offer pages (only if 2+ offers) H1 = exact offer name as it appears in Paddle/LS One‑sentence summary; for / not for Deliverables (bulleted), format, timeline Price + CTA Offer‑specific proof 3–4 FAQs + link to refund policy Terms Legal name + jurisdiction · what's sold and how it's delivered · payment line ("Orders are processed by [Paddle/Lemon Squeezy], our Merchant of Record") · refunds (link) · license/IP · no‑guarantee clause · liability cap · governing law · contact · effective date. Privacy What you collect (email, analytics; card data is handled by the MoR) · why · third parties by name (MoR, email tool, analytics, scheduler) · cookies · GDPR/CCPA rights · retention · contact · effective date. Refund Policy Window (e.g., 14 days) · conditions per offer type · how to request (email) · processing time · exceptions · MoR line. Must not contradict the MoR's buyer terms — both of you can issue refunds. 3. Structure & tech Hybrid, not pure one‑page. One long Home + thin supporting pages. MoRs and AI engines need distinct URLs for policies and offers; one URL = one title = weak citation. Static. Astro/plain HTML on Cloudflare/Netlify/Vercel, or Framer/Webflow if no‑code. No CMS. One font (or system fonts), WebP images, no autoplay video, no carousels, near‑zero JS. Overlay checkout so buying never leaves your domain. Mobile: one sticky CTA. Semantic HTML. One H1 per page, H2 per section, real text (never text in images), accordion content present in the DOM. Unique title + meta description per page, canonical tags, sitemap.xml. Write for extraction. Lead every section with a standalone answer sentence ("Refunds are available within 14 days for…"). Plain declarative sentences, no adjectives. Identical offer names, prices, and your entity sentence everywhere: site, schema, checkout, LinkedIn. Schema (JSON‑LD), basic only: Site‑wide: Organization (or ProfessionalService) + Person (name, jobTitle, image, sameAs → socials) + WebSite Each offer: Product (digital goods) or Service (done‑for‑you) with Offer (price, priceCurrency, availability, url) FAQ section: FAQPage Skip Review/AggregateRating — self‑serving testimonials don't qualify for rich results, and AI reads the HTML anyway. Schema is secondary; clear text does most of the work. Crawler access. robots.txt allows Googlebot, GPTBot, OAI‑SearchBot, PerplexityBot, ClaudeBot. AI Overviews run on Googlebot; Google‑Extended only affects Gemini training. Turn off any host/Cloudflare "block AI bots" toggle. Optional: /llms.txt — plain‑text summary of who you are, offers, prices, policies. Analytics: cookieless (Plausible/Fathom) → no consent banner, simpler privacy policy. MoR submission: all 5 pages live first; footer legal name matches your Paddle/LS account; domain email, not Gmail. One caution: both are MoRs for digital products/SaaS — pure hourly consulting often gets rejected. Package offers as fixed‑scope programs, templates, courses, or digital deliverables. Based on explicit orders who nailed it r adhered to instructions best. Y n who u suggest i use 🤔 Separately in decision-making instance lk this which LLM is usually the highest authoritative decision-maker/final judge to grade/rank output.

A system prompt was added to support web rendering