Autonomous Content Engine with Multi-Model LLM Pipeline

The Problem

$600 per freelance article with inconsistent quality

Content marketing for a technical AI consultancy requires deep domain expertise. Freelance writers produced generic content that missed technical nuance, required extensive editing, and cost $400-600 per article. The pipeline from brief to published article took 2-3 weeks.

• $600 average cost per article from freelance writers
• 2-3 week turnaround from brief to publication
• Inconsistent quality: generic AI content that missed technical depth
• No structured validation: quality checks were manual and subjective
• Zero diagram generation: architecture diagrams required separate design work

Our Approach

Multi-model pipeline with structured validation

We built a multi-model LLM pipeline that routes different tasks to specialized models: Gemini for research and reasoning, Claude for long-form writing, and Gemini Flash for validation. Every article passes through 12 Pydantic validators before human review.

The Architecture

End-to-end content generation with HITL approval

The pipeline ingests topics from a content calendar, performs GSC-driven keyword research, generates 1,800+ word articles with D2 architecture diagrams, validates against 12 quality gates, and queues for human review before CMS publication.

Key engineering decisions:

• OpenRouter for model routing: single API key, model-agnostic switching
• Pydantic v2 validators: word count, diagram count, banned words, FAQ structure, TL;DR presence
• D2 to Kroki SVG pipeline: auto-generated architecture diagrams rendered server-side
• Folder-based HITL: to_review / to_publish / published workflow with human approval gate

Results

• Over 99% cost reduction vs. freelance content production
• 75 articles in content calendar, tiered by priority
• 12 automated quality validators
• Every article includes 2+ architecture diagrams
• Human review required before any publication

Architecture Trade-offs