A comprehensive, system-level prompt engineered for AI coding agents (Claude, ChatGPT, Gemini, Copilot, Cursor, etc.) to build production-ready, SEO-optimized websites with minimal human intervention.
You are not just a web developer. This prompt activates a multi-role persona stack — Technical SEO Engineer, Information Architect, Core Web Vitals Expert, Schema.org Specialist, Content Strategist, and more — all working in parallel toward one goal: maximum ranking probability across search engines.
- What Makes This Different
- What Was Changed & Why
- Why This Prompt Exists
- Persona Stack
- Key Capabilities
- How It Works
- Who Should Use This
- Usage Instructions
- What It Covers
- Versioning
Most SEO prompts on the market fall into one of two camps:
| Type | Problem |
|---|---|
| Copy-paste checklist prompts | They list SEO tasks without enforcing execution. The AI nods along, generates generic metadata, and moves on. No real optimization happens. |
| Overly narrow prompts | They focus on one aspect (e.g., "write meta descriptions") and ignore the full lifecycle — content architecture, Core Web Vitals, schema, internal linking, crawl budget, post-build audit. |
This prompt is different because it is a system-level operational protocol, not a suggestion list.
| Feature | Typical Prompts | SEO Master Prompt |
|---|---|---|
| Persona activation | "Act as an SEO expert" (vague) | 12 distinct, tightly scoped personas with explicit responsibilities |
| Research depth | "Do keyword research" (one line) | 17 keyword/ intent types automatically discovered, with prioritization logic (low competition + high volume + growing trend + easy rank) |
| Live data requirement | Relies on cached AI knowledge | Mandates live web research before any code is written |
| Code generation | Suggests best practices | Enforces technical implementation: schema, robots, sitemaps, Core Web Vitals, WCAG, CSP, image optimization |
| Post-build audit | Often skipped entirely | Mandatory automated audit across 24 checkpoints; repeats until scores hit 95+ |
| Dashboard output | Absent | Final SEO dashboard with quantified scores across all dimensions |
| Self-correction loop | None | If any score is below 90, the agent must continue improving until all targets are met |
This prompt has been refined through multiple iterations. The key evolutions:
Always perform live web research before development.
Use the latest available data to identify trending, growing,
low-competition, high-search-volume keyword opportunities.
Do not rely on cached knowledge.
Why: AI training data is months (sometimes years) old. Search trends shift weekly. The prompt now forces the agent to fetch real-time data before generating anything, ensuring the SEO strategy is current — not a snapshot of last year's SERP.
Beyond Google, the prompt explicitly targets: Bing, Perplexity, ChatGPT Search, Gemini Search, Claude Search, and future AI-driven search systems.
Why: The search landscape is fragmenting. AI-native answer engines use different ranking signals than traditional crawlers. Optimizing only for Google leaves traffic on the table.
Instead of a single "SEO expert" instruction, the prompt activates 12 distinct personas with overlapping but distinct responsibilities.
Why: SEO is not one discipline. A single persona produces shallow output. A coordinated team of specialized roles produces comprehensive, production-grade results.
Technical SEO Score: 95+/100
On-page SEO: 95+/100
Performance: 95+/100
If any score is below 90, automatically continue improving.
Why: Vague instructions like "make it good" produce unpredictable results. Hard numerical targets create an objective completion criterion that both the AI and the human can verify.
Why: Most AI-generated code looks correct but fails on validation (missing canonical, broken schema, skipped ALT tags, duplicate titles). The audit catches what the generation phase misses and fixes it automatically.
Why: Without a summary, the human has to manually inspect 20+ dimensions. The dashboard gives instant visibility into every critical metric, making verification a 10-second read.
Building an SEO-optimized website requires expertise across multiple domains — technical SEO, content strategy, semantic architecture, performance engineering, accessibility, structured data, conversion optimization. No single developer (and certainly no generic AI prompt) covers all of these adequately.
Most developers:
- Write meta descriptions but forget canonical URLs
- Add schema markup but break Core Web Vitals
- Optimize images but miss heading hierarchy
- Build mobile-responsive layouts but fail WCAG accessibility
- Generate sitemaps but forget image sitemaps
- Target keywords but ignore search intent
This prompt encodes decades of combined SEO engineering experience into a single, executable instruction set. It bridges the gap between:
- What an SEO specialist knows but can't always code
- What a developer can build but doesn't know to optimize
- What an AI agent can do but isn't instructed to do comprehensively
The result: a website that is production-ready, technically optimized, and genuinely competitive in search rankings — generated in a single pass.
| Role | Why |
|---|---|
| Indie hackers & founders | Ship SEO-optimized landing pages, SaaS sites, and MVPs without hiring an SEO specialist |
| SEO agencies | Use as a standardization layer — every site they build meets a baseline technical SEO score of 95+ |
| Web developers | Eliminate the gap between "it works" and "it ranks" |
| AI power users | Get production-grade output from AI agents instead of mediocre, uncooked results |
| Content teams | Ensure every published page satisfies search intent, topical authority, and semantic coverage |
The prompt simultaneously activates these personas. Each has defined responsibilities that intersect and reinforce each other:
| # | Persona | Core Responsibility |
|---|---|---|
| 1 | Senior Technical SEO Engineer | Crawl budget, indexability, site architecture, technical audits |
| 2 | Google Search Quality Expert | E-E-A-T signals, search quality rater guidelines compliance |
| 3 | Information Architect | Content hierarchy, topical silos, semantic relationships |
| 4 | Semantic SEO Specialist | Entity optimization, NLP keyword coverage, topical clusters |
| 5 | Topical Authority Strategist | Pillar pages, content hubs, internal link topology |
| 6 | Content Strategist | Search intent satisfaction, content depth, user value |
| 7 | UX Engineer | Mobile-first design, responsiveness, conversion flow |
| 8 | Core Web Vitals Expert | LCP < 2.5s, CLS < 0.1, INP < 200ms |
| 9 | Accessibility Expert | WCAG AA+, keyboard nav, ARIA, screen readers, contrast |
| 10 | Schema.org Expert | Structured data, rich results eligibility, schema validation |
| 11 | Crawl Budget Optimizer | Robots.txt, sitemap strategy, crawl efficiency |
| 12 | Conversion Rate Optimization Specialist | CTAs, trust signals, testimonials, form optimization |
| 13 | Frontend Performance Engineer | Code splitting, tree shaking, critical CSS, lazy loading |
| 14 | Backend Performance Engineer | Compression, caching, CDN strategy, server response times |
The prompt doesn't ask for keywords — it discovers them:
- Primary, secondary, semantic, NLP, LSI keywords
- Long-tail, question-based, emerging, trending, seasonal keywords
- Intent clusters: informational, commercial, transactional, navigational
- Low-competition, high-search-volume, growing-trend opportunities
- Competitor keyword gaps
| Schema | Auto-generated |
|---|---|
| Breadcrumb | ✅ Always |
| Organization | ✅ Always |
| Website | ✅ Always |
| FAQ | ✅ Always |
| Article | ✅ When blog/content |
| Product | ✅ When eCommerce |
| SoftwareApplication | ✅ When SaaS |
| Review | ✅ When reviews exist |
| LocalBusiness | ✅ When local |
| Video | ✅ When video content |
| Event | ✅ When events |
| Course | ✅ When educational |
| HowTo | ✅ When instructional |
| Recipe | ✅ When food/recipe |
robots.txt— optimized crawl directivessitemap.xml— proper priority/ changefreqimage-sitemap.xml— all images includedvideo-sitemap.xml— when video presentrss.xml— when blog/ content updates
- SEO-optimized filenames (descriptive, keyword-rich)
- ALT tags (every image, context-aware)
- Title attributes
- Lazy loading (
loading="lazy") - Modern formats (WebP, AVIF)
- Responsive srcset
| Metric | Target |
|---|---|
| LCP (Largest Contentful Paint) | < 2.5 seconds |
| CLS (Cumulative Layout Shift) | < 0.1 |
| INP (Interaction to Next Paint) | < 200ms |
| Lighthouse Performance | 95+ |
| Standard | Requirement |
|---|---|
| WCAG Level | AA+ |
| Keyboard navigation | Full |
| ARIA labels | Every interactive element |
| Screen reader | Semantic HTML + ARIA live regions |
| Color contrast | 4.5:1 (normal), 3:1 (large) |
| Focus states | Visible, logical tab order |
Every site undergoes an automated audit checking:
- ✅ Metadata completeness
- ✅ Canonical URLs
- ✅ Structured Data validity
- ✅ Sitemap correctness
- ✅ Robots.txt directives
- ✅ Heading hierarchy
- ✅ Internal link structure
- ✅ Broken links (internal + external)
- ✅ Duplicate titles
- ✅ Duplicate descriptions
- ✅ Missing ALT tags
- ✅ Core Web Vitals (simulated)
- ✅ Accessibility (WCAG AA+)
- ✅ Mobile friendliness
- ✅ Indexability (noindex/ meta robots)
- ✅ Crawlability
- ✅ Semantic coverage
- ✅ Entity coverage
- ✅ Keyword coverage
- ✅ Rich Results eligibility
- ✅ Open Graph tags
- ✅ Twitter Card tags
- ✅ Page speed
- ✅ Security headers
┌─────────────────────────────────────────────────────────┐
│ INPUT: Website Idea │
│ (topic, domain, or product) │
└─────────────────────┬───────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────┐
│ PHASE 1: LIVE RESEARCH │
│ • Real-time keyword discovery (web search) │
│ • Competitor analysis │
│ • Trend identification │
│ • Search intent mapping │
└─────────────────────┬───────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────┐
│ PHASE 2: ARCHITECTURE │
│ • Topical authority mapping │
│ • Information architecture design │
│ • Content silo planning │
│ • Internal link topology │
│ • Schema strategy │
└─────────────────────┬───────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────┐
│ PHASE 3: DEVELOPMENT │
│ • HTML/CSS/JS generation │
│ • Schema markup injection │
│ • Image optimization │
│ • Performance optimization │
│ • Accessibility implementation │
│ • Security hardening │
└─────────────────────┬───────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────┐
│ PHASE 4: SEO AUDIT │
│ • 24-point automated audit │
│ • Score computation │
│ • Issue detection │
│ • Automatic fix loop │
└─────────────────────┬───────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────┐
│ PHASE 5: DASHBOARD │
│ • Comprehensive score summary │
│ • Pass/fail per dimension │
│ • Remaining improvements (if any) │
│ • Production-ready confirmation │
└─────────────────────────────────────────────────────────┘
- Copy the contents of
SEO_Master_Prompt.md - Paste it as the system prompt or initial instruction
- Follow with your website request:
- "Build a landing page for [topic]"
- "Create a SaaS website for [product]"
- "Generate an eCommerce product page for [item]"
- "Build a blog with topical authority on [niche]"
Most AI agents accept system prompts that persist for the entire session. This gives the best results as the prompt shapes every response.
Include SEO_Master_Prompt.md in your project's root. Reference it in cursor rules, agent instructions, or .claude config files.
| Category | Coverage |
|---|---|
| Technical SEO | Crawl budget, indexability, canonicalization, robots.txt, sitemaps, redirects, pagination, URL structure |
| On-Page SEO | Title tags, meta descriptions, headings, content optimization, keyword density, semantic relevance |
| Content Strategy | Topical authority, content clusters, pillar pages, search intent, entity optimization |
| Structured Data | 14+ schema types, JSON-LD, rich results, schema validation |
| Performance | Core Web Vitals, Lighthouse 95+, code splitting, lazy loading, caching, CDN |
| Accessibility | WCAG AA+, ARIA, keyboard nav, screen readers, contrast, focus management |
| Security | HTTPS-ready, CSP, input sanitization, secure headers |
| Mobile | Responsive design, touch targets, mobile-first CSS, viewport optimization |
| Conversion | CTA strategy, trust signals, testimonials, social proof, form UX |
| International SEO | hreflang (when multilingual), multi-region support |
| Image Optimization | ALT tags, filenames, WebP/AVIF, srcset, lazy loading |
| Post-Build Audit | 24 automated checks, self-correction loop, quantified scoring |
| Version | Date | Changes |
|---|---|---|
| v1.0 | June 2026 | Initial release — 12-persona stack, 17 keyword types, 14 schemas, 24-point audit, dashboard output |
| v1.1 | June 2026 | Added live web research mandate before development. Expanded search engine targets to AI-native platforms (Perplexity, ChatGPT Search, Gemini Search, Claude Search). |
MIT License — See LICENSE file.
Free to use, modify, and distribute. No attribution required (MIT standard).
Built for anyone who wants their website to rank — not just look good.