HHE SEO/AEO Content Pipeline

Internal Technical Briefing — March 10, 2026

Design Approved — Awaiting Client Inputs

Project Overview

What: Automated SEO/AEO content generation pipeline for Holistic Health Educators (HHE) and the Lily Fields brand.

Why: Competitor analysis shows top holistic certification programs are poorly optimized for AI search engines — first-mover advantage in AEO positioning.

Scale: 60-page backlog across 5 page types. Weekly publishing cadence after an initial rollout of 15–20 core pages.

Voice: All content written in Karen Hurd's brand voice via RAG retrieval from the existing holistic health knowledge base.

System Architecture

ExternalStorageCloudflare WorkerDashboardWeb UIAPI RoutesGeneratorQA ScorerRAG ClientD1 SQLiteR2 Blobsproxy-chatChromaDBGemini APIFlash + Pro
LayerTechnologyPurpose
ComputeCloudflare Worker (TypeScript)API + dashboard serving
DatabaseCloudflare D1 (SQLite)Topics, pages, QA scores, audit logs
File StorageCloudflare R2Draft HTML/markdown content
Content GenerationGoogle Gemini FlashFast draft generation
Quality ScoringGoogle Gemini ProDeep QA gate enforcement
Brand Voice RAGExisting proxy-chat HTTP APIChromaDB knowledge base
DashboardVanilla HTML/CSS/JSNo build step needed

Generation Pipeline Flow

D1 + R2Gemini ProGemini FlashRAG (proxy-chat)CF WorkerDashboardD1 + R2Gemini ProGemini FlashRAG (proxy-chat)CF WorkerDashboardPOST /api/generate/:topicIdCreate run_log (started)202 { run_id }POST /api/rag/query(brand voice)Karen Hurd documentsGenerate draft(Flash model)Raw draft HTMLStore draft in R2 + D1QA score draft(Pro model)Score resultsStore QA in qa_runsGET /api/runs/:runId{ status: done, score: 82 }

Each topic goes through async generation: RAG retrieval → Gemini Flash draft → Gemini Pro QA scoring → D1/R2 storage. The dashboard polls for completion and shows results with full audit trail.

QA Scoring System

Required Gates (Binary Pass/Fail)

GatePass Criteria
H1 countExactly 1
H2 count>= 2
Meta descriptionPresent, 120–160 chars
FAQ block>= 3 Q&A pairs
ComplianceNo banned health claims

Scored Checks

30%20%15%15%10%10%AEO Score Composition (100 pts)Brand VoiceContent DepthInternal LinksBLUF IntroSchemaCTA Block
Pass threshold: >= 75 / 100

Page Types

1Program Page

Course overview, curriculum, outcomes, enrollment CTA

2Comparison Page

Feature table, pros/cons, recommendation

3Alternatives Page

Top alternatives list, why-us differentiation

4Career Page

Career path, requirements, salary, how-to-start

5Accreditation Page

Details, standards, credibility markers

Topic Status Flow

POST /generateScore >= 75Score < 75Pipeline failureRetryManual approveRe-generateRetryPhase 2Phase 2pendinggeneratingqa_passqa_failerrorapprovedscheduledpublished

Topics progress through the pipeline: pendinggeneratingqa_pass / qa_failapproved. In Phase 2, the flow extends to scheduledpublished via WordPress CMS integration. Failed QA runs can be retried or re-generated.

Implementation Timeline

2026-03-112026-03-132026-03-152026-03-172026-03-192026-03-212026-03-232026-03-252026-03-272026-03-292026-03-312026-04-012026-04-032026-04-05Scaffold + SchemaRouter + Topics APIRAG + Gemini ServicesQA ScorerTemplates + GeneratorAPI RoutesDashboard + E2EWordPress CMS AdapterCalendar + SchedulingAuth + Bulk OpsPhase 1: Prove PipelinePhase 2: Full StackImplementation Timeline

Phase 1 — "Prove the Pipeline": 7 implementation chunks, 12 tasks, TDD approach. Scaffold through dashboard + end-to-end testing.

Phase 2 — "Full Stack": WordPress CMS adapter, calendar scheduling, Cloudflare Access auth, and bulk operations.

Outstanding Client Inputs

The following decisions are needed from Blake before implementation can begin. Submit responses via the decision form.

  1. 60-page topic plan and rollout order
  2. Gold-standard examples per page type
  3. Acceptance checklist and scoring rubric
  4. Cloudflare account access
  5. Gemini API key
  6. WordPress site details (Phase 2 prep)