完整的 AI SEO 实操指南:从零到三个月内获得 460 万次展示
The complete AI SEO playbook: from zero to 4.6M impressions in 3 months

原始链接: https://github.com/TraceCohenTech/ai-seo-playbook

本手册概述了一套经实战检验的方法论,用于构建高排名的 AI 内容引擎。ValueAddVC.com 曾利用该方法在短短三个月内将月度展示量从 60.4 万提升至 462 万。 该系统不仅仅依赖简单的提示词,而是充当内容的“操作系统”,将多模型 AI 智能体编排(Opus、Sonnet、Haiku)与严格的反馈回路相结合。其框架基于四大支柱: * **GSC 反馈回路:** 通过自动化诊断脚本分析 Google Search Console 数据,识别标题重写机会、关键词缺口及关键词蚕食集群。 * **质量与安全系统:** 设有九道“发布关卡”和自动化黑名单,剔除 AI 生成的“模板化特征”,确保内容保持类人质感与高质量。 * **技术优化:** 包含 CI/CD 自动化、内部链接图管理、Schema 验证,以及通过 WebSub 和索引 API 实现即时索引。 * **增长迭代:** 通过关键词预判、旧页面“刷新滴灌”及内容审计(删除/合并/更新/提升)等策略,使引擎转变为自我改进的循环系统。 本工具包提供 14 个诊断脚本、结构化数据 Schema 以及 GitHub Actions,旨在实现全流程自动化,使用户能够复制一套以数据驱动内容表现的高增长 SEO 策略。

Hacker News 最新 | 过往 | 评论 | 提问 | 展示 | 招聘 | 提交 登录 完整的 AI SEO 实战手册:三个月内从零到 460 万次展示量 (github.com/tracecohentech) 由 t_cohen 发布于 1 小时前 | 5 分 | 隐藏 | 过往 | 收藏 | 3 条评论 | 帮助 oefrha 4 分钟前 | 下一条 [–] ValueAddVC 恬不知耻地炫耀自己如何用垃圾内容给互联网制造负面价值。经典。(我知道,评论这个就是在为他们的 SEO 努力做贡献。) 回复 an0malous 9 分钟前 | 上一条 [–] 这就是为什么互联网在五年内将彻底沦为一坨垃圾的原因。当然,发帖的还是个风投(VC)。 回复 Gualdrapo 3 分钟前 | 父评论 [–] 你指的“将”是什么意思?我们难道还活在 2021 年吗? 回复 指南 | 常见问题 | 列表 | API | 安全 | 法律 | 申请 YC | 联系 搜索:
相关文章

原文

License: MIT Node 20+ PRs Welcome GitHub stars

The complete playbook for building an AI-powered content engine that actually ranks — from zero to 4.6M impressions in 3 months.

This is the methodology, the toolkit, and the hard-won lessons from building a content engine on ValueAddVC.com using AI agents, GSC feedback loops, and automated quality gates. 14 diagnostic scripts, 9 battle-tested configs (safety guards, agent orchestration, quality gates, anti-AI detection), structured data schemas, and CI automation — everything you need to replicate the system.

Not theory. Not prompts. The actual operating system behind a site that went from 604K to 4.62M monthly impressions.

Built by Trace Cohen at ValueAddVC.com.


What This Playbook Covers

  1. The Content Engine — AI agent orchestration (multi-model pipelines: Opus/Fable for planning, Sonnet for writing, Haiku for grunt work), 5-format content rotation, voice training, anti-AI fingerprint detection
  2. The GSC Feedback Loop — Weekly automated reports, title rewrite candidates, cannibalization detection, query gap mining, striking distance optimization
  3. The Quality System — 9 publish gates, template phrase blocklists, source verification, fact-checking, structured data validation
  4. The Safety Layer — Repo locks, rebase guards, build cost control ([nobuild] tags, deploy-tick), self-healing heartbeats, content writer isolation from git
  5. The Growth Loop — Keyword anticipation (publish before demand spikes), living page refreshes, internal link graph optimization, news sitemap + WebSub for instant crawling

flowchart LR
    GSC[Google Search Console API] --> Report[Weekly Report]
    Report --> Triage{Triage}
    Triage --> Rewrite[Title Rewrites]
    Triage --> Merge[Merge Cannibals]
    Triage --> Refresh[Refresh Stale Pages]
    Triage --> Kill[Kill Dead Weight]
    
    Content[Content Pipeline] --> Gates[Quality Gates]
    Gates --> |Pass| Publish[Publish]
    Gates --> |Fail| Fix[Fix & Retry]
    
    Publish --> GSC
    Rewrite --> GSC
    Merge --> GSC
    Refresh --> GSC

    style GSC fill:#0d7377,color:#fff
    style Gates fill:#0d7377,color:#fff
    style Publish fill:#15803d,color:#fff
    style Kill fill:#b91c1c,color:#fff
Loading

The feedback loop: GSC data feeds diagnostic scripts → scripts surface what needs fixing → AI agents make the fixes through quality gates → improved rankings produce better GSC data → repeat. Every week the system gets smarter.


Script What It Does
gsc-rewrite-candidates.mjs Finds title rewrite opportunities from GSC data — pages ranking position 4–20 with high impressions but low CTR
template-detector.mjs Scans your content for AI template fingerprints — the repeated phrases that signal scaled-content-abuse to Google
cannibalization-detector.mjs Finds pages on your site competing for the same queries, splitting authority and ranking worse than one consolidated page would
weekly-report.mjs Generates a weekly SEO performance report with trending queries, dropping pages, CTR triage candidates, and query monopolies
orphan-finder.mjs Finds pages with zero inbound internal links — invisible to Google's link-graph crawler
content-audit.mjs Scores every page into KILL / MERGE / UPDATE / PROMOTE / KEEP buckets based on GSC data + content quality
redirect-checker.mjs Finds URLs in your sitemap that return 301/302/308 instead of 200 — these break GSC validation and waste crawl budget
refresh-tracker.mjs Identifies high-traffic pages that haven't been updated recently — candidates for the "refresh drip" strategy
query-gap-miner.mjs The retroactive keyword discovery engine — finds queries with real demand where you have no dedicated page. Google is telling you what to write.
striking-distance.mjs Finds pages ranking position 5-20 with real impressions — the cheapest wins in SEO. Estimates click gain if improved.
rewrite-measurer.mjs Before/after tracking for title rewrites. Take a baseline, make changes, measure impact 2-4 weeks later.
websub-ping.mjs Notifies Google's hub that your feeds changed — triggers immediate crawl instead of waiting hours. Run after every publish.
indexing-submitter.mjs Submits URLs to Google's Indexing API for near-instant crawling. 200 URLs/day quota.
broken-link-checker.mjs Scans all content for outbound links and checks for 404s, timeouts, and redirect chains. Exits non-zero for CI.
File Purpose
format-rotation.json The 5-format content system: Deep Explainer, News Analysis, Ranked List, Question-Led, Contrarian Take — with per-format word counts, chart requirements, and selection weights
quality-gates.json Publish gate rules: cannibalization check, source URL verification, template phrase detection, shared closer detection, typecheck
anti-ai-rules.json The complete blocklist of AI template phrases + style rules for making AI content sound human
refresh-rules.json Rules for the refresh drip strategy — staleness thresholds by content type, refresh triggers, and a refresh checklist
keyword-anticipation.json Event calendar methodology — publish content before IPOs, earnings, funding rounds, regulations so you're ranked when demand spikes
health-checks.json Live-site health checks: leaked template variables, broken OG images, injected ad links, thin content, dead pages
content-pipeline-guards.json Safety guards: repo locks, rebase guards, cannibalization checks, build cost control, self-healing heartbeats
agent-orchestration.json Multi-model AI pipeline rules: Opus/Fable for planning, Sonnet for writing, Haiku for mechanical tasks. Max 3 concurrent agents.

Schema Examples (/schemas)

File Schema Type
article-with-author.json Article + Person author entity (the E-E-A-T foundation)
faq-page.json FAQPage for blog posts — drives FAQ rich results
item-list.json ItemList for ranking/comparison pages — the format sponsors want
news-article.json NewsArticle + news sitemap template for real-time content
  • sitemap.ts — Next.js dynamic sitemap with honest lastmod dates
  • news-sitemap.ts — 48-hour rolling news sitemap for Google News/Discover
  • internal-link-component.tsx — React component for related posts + a build-time internal link inserter
  • vercel-ignore.sh — Build skip logic for Vercel: [nobuild] tags, content-only detection, deploy-tick pattern (saves $$$)

Every script has a sample output file so you can see what to expect before running anything:

  • weekly-report.json — Full weekly report with trending queries, dropping pages, CTR triage
  • rewrite-candidates.json — Title rewrite opportunities with per-query diagnosis
  • content-audit.json — KILL/MERGE/UPDATE/PROMOTE/KEEP bucket assignments
  • cannibal-clusters.json — Cannibalization clusters with wasted impression estimates
  • template-scan.json — AI fingerprint scan with per-file phrase locations
  • orphan-pages.json — Orphan, low-link, and dead-end page reports
  • setup-gsc.md — Step-by-step Google Search Console API setup (local auth + service account for CI)

Automation (.github/workflows)

  • weekly-seo-report.yml — GitHub Action that runs the weekly report every Sunday, commits results, and optionally creates a GitHub issue with the summary

# Clone the repo
git clone https://github.com/TraceCohenTech/ai-seo-playbook.git
cd ai-seo-playbook

# Install dependencies
npm install

# Set up Google Search Console API access
# (requires a Google Cloud project with Search Console API enabled)
gcloud auth application-default login \
  --scopes=https://www.googleapis.com/auth/webmasters.readonly

# Find title rewrite opportunities
npm run rewrite-candidates -- --site sc-domain:yoursite.com

# Scan for AI template fingerprints
npm run template-scan -- --dir ./your-content-directory

# Find cannibalization clusters
npm run find-cannibals -- --site sc-domain:yoursite.com

# Run a full content audit
npm run content-audit -- --site sc-domain:yoursite.com --dir ./your-content-directory

# Find orphan pages (no internal links)
npm run find-orphans -- --dir ./your-content-directory

# Generate weekly report
npm run weekly-report -- --site sc-domain:yoursite.com

# Discover keywords you're already ranking for but have no page targeting
npm run query-gaps -- --site sc-domain:yoursite.com --dir ./your-content-directory

# Find "almost page 1" pages where a small nudge = big click gains
npm run striking-distance -- --site sc-domain:yoursite.com

# Find stale pages that need refreshing
npm run refresh-tracker -- --site sc-domain:yoursite.com --dir ./your-content-directory

# Check for redirect problems in your sitemap
npm run check-redirects -- --site sc-domain:yoursite.com --sitemap https://yoursite.com/sitemap.xml

# Ping Google to crawl your updated feeds immediately
npm run websub-ping -- --feeds https://yoursite.com/sitemap.xml,https://yoursite.com/feed.xml

New to the GSC API? See docs/setup-gsc.md for a step-by-step setup guide.


These tools are one half of the system. The methodology — why these specific metrics matter, how to interpret the results, and how to build the feedback loop that makes your content engine self-improving — is in the full guide:

The AI SEO Playbook: How I Used AI to Build a Content Engine That Hit 4.6M Impressions in 3 Months

The guide covers:

  • Building the content engine (architecture, voice training, format rotation)
  • The GSC reckoning (the AI-overview discovery, title rewrites, cannibalization)
  • The iteration loop (keyword anticipation, living pages, technical SEO bugs)
  • The system (quality gates, weekly reviews, cost control)

These tools were built and refined on ValueAddVC.com over 3 months:

Metric Week 1 (May '26) Now (Aug '26)
3-Month Impressions 4.62M
3-Month Clicks 17.3K
Daily Clicks (peak) ~50 854
Average Position 12+ 7.5
CTR 0.93% 0.4%
Posts Audited 480 960+
Title Rewrites 0 92
Cannibalization Clusters Fixed 0 21
Template Phrases Purged 500+ 0
Orphan Pages Linked 0 191

*CTR is 0.4% because impressions grew ~8x — largely from AI-overview citations (GEO traffic) that don't produce clicks by nature. Human-intent CTR improved: ranked lists hit 6.8%, question-led posts hit 3.2%. The growth curve is near-vertical: Aug 13 alone hit 127K impressions and 854 clicks.


How to Set Up the Weekly Cron

Option 1: GitHub Actions (recommended)

  1. Create a Google Cloud service account with Search Console API access
  2. Add the service account JSON as a GitHub secret named GSC_CREDENTIALS
  3. Set the repository variable GSC_SITE to your GSC property (e.g., sc-domain:yoursite.com)
  4. Set CONTENT_DIR to your content directory path (e.g., ./src/app/blog)
  5. Optionally set CREATE_ISSUES to true for weekly GitHub issue summaries
  6. The workflow runs every Sunday at 9:30 AM ET automatically

Option 2: Local cron (macOS launchd)

# Create a plist in ~/Library/LaunchAgents/
# Schedule: every Sunday at 9:30 AM
# Script runs: node scripts/weekly-report.mjs --site sc-domain:yoursite.com
# Commits results to git

Option 3: Any CI/CD system

The scripts are standalone Node.js — run them anywhere you can install googleapis and authenticate with Google Cloud.


ai-seo-playbook/
├── scripts/              # 14 diagnostic & tracking scripts
│   ├── weekly-report.mjs            # Weekly GSC performance report
│   ├── gsc-rewrite-candidates.mjs   # Find title rewrite opportunities
│   ├── rewrite-measurer.mjs         # Before/after rewrite tracking
│   ├── query-gap-miner.mjs          # Retroactive keyword discovery
│   ├── striking-distance.mjs        # Position 5-20 opportunities
│   ├── template-detector.mjs        # Scan for AI template phrases
│   ├── cannibalization-detector.mjs  # Find competing pages
│   ├── content-audit.mjs            # KILL/MERGE/UPDATE/PROMOTE scoring
│   ├── orphan-finder.mjs            # Find unlinked pages
│   ├── refresh-tracker.mjs          # Stale page detection
│   ├── redirect-checker.mjs         # Sitemap redirect problems
│   ├── broken-link-checker.mjs      # 404s and dead outbound links
│   ├── websub-ping.mjs              # Notify Google of feed changes
│   └── indexing-submitter.mjs       # Google Indexing API submissions
├── config/               # Quality gates, format system, anti-AI rules
├── schemas/              # JSON-LD structured data examples
├── examples/             # Next.js sitemaps + React components
├── samples/              # Example output from every script
├── docs/                 # Setup guides
└── .github/workflows/    # Weekly automated report CI

Found a template phrase pattern that should be in the blocklist? A better heuristic for the content audit scorer? See CONTRIBUTING.md for how to submit changes.

There's even a dedicated issue template for submitting new template phrases — the blocklist is never complete.


This toolkit was built and battle-tested on ValueAddVC.com — a venture capital content platform that went from 604K monthly impressions to 4.62M in 3 months using these exact scripts and methodology.

May 2026 August 2026
Daily clicks ~50 854 (peak)
Position 12+ 7.5
Template phrases 500+ 0
Orphan pages 191 0

The full methodology is in the companion guide: The AI SEO Playbook


MIT


Built by Trace Cohen · ValueAddVC.com · [email protected]

联系我们 contact @ memedata.com