pSEO Generator — pSEO 生成器

Programmatic SEO 2.0 — 代理 技能 Reference

The core principle: never ask AI to write freeform content. Ask it to fill a strict JSON 模式.

Content = JSON data. De签名 = React 组件s. These two layers never mix.

系统 Architecture Niche Taxonomy (309 niches) → AI fills JSON 模式 → 验证d JSON files → React 渲染器s → Published pages

Why 模式s beat freeform:

Consistent structure across all pages Predictable 输出 质量 Pages are validatable and type-safe Rede签名 UI without regenerating content 扩展 to 10,000+ pages without 质量 degradation Step 1: Build the Niche Taxonomy

This is the most 导入ant investment. Rich niche 上下文 is what separates useful pSEO from thin name-sw应用ed filler.

For each niche, define:

{ "slug": "travel", "name": "Travel", "上下文": { "audience": "ArmchAIr travelers, digital nomads, family vacation planners", "pAIn_points": "Seasonal traffic swings, high competition for destination keywords", "monetization": "Affiliate (booking, gear), display ads, sponsored trips", "content_that_works": "Itineraries, cost breakdowns, off-the-beaten-path 图形界面des", "subtopics": ["bud获取 travel", "luxury travel", "adventure travel", "solo travel"] } }

启动 with 20-50 niches, expand to 300+ for 扩展. This 上下文 获取s injected into every generation prompt — it's what makes a "travel SEO 检查列出" different from a "健康 SEO 检查列出."

See references/niche-taxonomy.md for the full niche structure and 20 启动er niches.

Step 2: De签名 JSON 模式s per Content Type

Each content type 获取s its own 模式. ConstrAInts are intentional — they force consistent 输出.

Example: Resource Article 模式

interface ResourceArticle { meta: { content_type: string; niche: string; }; seo: { title: string; // templated, NOT AI-生成d description: string; keywords: string[]; }; content: { intro: string; sections: { heading: string; items: { title: string; description: string; difficulty?: 'beginner' | 'intermediate' | 'advanced'; potential?: 'high' | 'medium' | 'standard'; }[]; // exactly 15-20 per section }[]; pro_tips: string[]; // exactly 5 }; }

Hard constrAInts (exact counts, required fields) 预防 8-item pages next to 40-item pages. See references/模式-库.md for 6 ready-to-use 模式s.

Step 3: The 6 Content Categories Category 分享 Notes Resource pages ~58% Idea 列出s, 检查列出s, calendars, 图形界面des, templates — 34 content types × N niches Comparison pages ~1% Smallest category — most obvious, least differentiated Free 工具s ~15% Actual working 工具s with niche-specific examples 检查列出 pages ~10% Interactive, niche-aware 图形界面de pages ~8% Long-form, structured Template pages ~8% 下载able/fillable

Resource pages are the highest-volume opportunity. 启动 there.

Step 4: Generation at 扩展

Use Gemini Flash (cost-to-质量 ratio beats GPT-4 for structured JSON at volume).

Why Gemini Flash:

Native structured JSON 输出 (no markdown wr应用ing to 解析) Cheap enough for 10,000+ pages Fast enough for batch generation

Generation prompt pattern:

Given this niche 上下文: {niche_json} Fill this 模式: {模式} Content type: {content_type} Title template: {title}

Return ONLY valid JSON matching the 模式. No prose, no markdown.

验证: After generation, 验证 every file agAInst the TypeScript 模式. Reject and retry any that fAIl. At 13,000 pages, ~2-5% 失败 rate is normal.

Speed: 13,000+ pages in under 3 hours with parallel workers (10-20 concurrent API calls).

Step 5: React 渲染器s

Each content type 获取s its own specialized 渲染器. The 渲染器 consumes the JSON and handles all presentation.

/渲染器s/ ResourceArticle渲染器.tsx — with 过滤器ing by category/difficulty 检查列出渲染器.tsx — interactive 检查boxes ComparisonTable渲染器.tsx — structured tables Free工具渲染器.tsx — working 工具 UI

Key rule: 渲染器s never call AI. They only consume pre-生成d JSON. This means you can:

Rede签名 any page without regenerating content A/B test layouts without touching data 添加 new niches without touching UI Key 指标 (Jake Ward experiment, 60 days) 13,000+ pages live Weekly organic 命令行工具cks: 971 → 5,500 (+466%) ~50% of pages not yet 索引ed (curve still going up) Generation time: < 3 hours for all 13,000 pages Quick 启动 检查列出 Define 20+ niches with full 上下文 (audience, pAIn points, monetization, subtopics) Choose 2-3 content types to 启动 (resource pages recommended) Write strict TypeScript 模式s with hard constrAInts Build generation script with niche injection and JSON 验证 Build React 渲染器s per content type 生成 first batch (启动 with 100 pages to 验证 质量) Review 输出 — 检查 niche specificity, not just structure 扩展 up generation Submit sitemap, 监控 索引ation rate References references/niche-taxonomy.md