A forked version of the original LLMO Optimization Analyzer - update to go beyond basic LLM scoring to provide comprehensive AI search optimization analysis.
This fork transforms the original LLMO analyzer into a complete AI search optimization suite that analyzes 15+ additional ranking factors beyond schema markup.
- Answer Engine Optimization (AEO): Detects direct answers, definitions, comparisons, and pros/cons sections
- Semantic Coverage: Tracks entities, topic depth, and unique technical terms
- Information Gain: Measures expert quotes, statistics, case studies, and original insights
- Content Structure: Analyzes hierarchy, scannability, chunk quality, and section independence
Original version detected basic schemas. This version:
- Detects 10+ schema types with pattern matching
- Validates schema claims against actual content
- Identifies schema conflicts (e.g., Article vs WebPage)
- Calculates schema completeness percentages
- Suggests missing high-impact schemas with ROI estimates
- E-E-A-T Scoring: Separate scores for Experience, Expertise, Authority, and Trust (0-100 scale)
- Topic Coverage Analysis: Measures content comprehensiveness
- Freshness Signals: Tracks temporal keywords and current year mentions
- Authority Markers: Analyzes citations, author bios, and trust signals
- Moved from simple 0-5 to comprehensive 0-100 scoring
- Multiple scoring dimensions instead of single score
- Shows potential score with optimizations
- Identifies primary limiting factors
- Multi-Intent Coverage: Evaluates if content satisfies multiple user intents
- Voice Search Optimization: Detects conversational patterns
- Question Type Coverage: Tracks what/why/how/when/where/who/which
- Answer Quality Rating: Partial vs complete vs comprehensive
- Estimated dwell time based on content depth
- Shareability score based on insights and data
- Reference value for bookmark-worthy content
- Return visit likelihood predictions
- Traffic Impact Estimates: Projects percentage increases
- Implementation Effort: Hours required for each improvement
- Industry Benchmarking: Compares schema count to industry standards
- Quick Wins Identification: Highlights easy, high-impact improvements
Original detected FAQs. This version:
- Identifies implied FAQ patterns ("how do", "what is", "when should")
- Generates FAQ suggestions from existing content
- Counts both explicit and implicit FAQ opportunities
New report sections include:
- 🚨 Critical Fixes with time/impact estimates
- 🎯 Quick Wins for immediate improvements
- Answer Engine Readiness scores
- LLM Parsing Optimization metrics
- Competitive Analysis insights
- Content Addition Recommendations
- Better content extraction with context preservation
- Improved passage weighting algorithms
- Enhanced pattern detection for 30+ LLM signals
- More accurate content type detection
| Feature | Original | Enhanced |
|---|---|---|
| Schema Types Detected | Basic | 10+ with validation |
| Scoring Scale | 0-5 | 0-100 with multiple dimensions |
| Content Signals | Limited | 30+ LLM optimization signals |
| FAQ Detection | Basic | Explicit + Implied patterns |
| E-E-A-T Analysis | No | Yes, with individual scores |
| ROI Estimates | No | Yes, with hours and impact |
| Industry Benchmarks | No | Yes, by content type |
| Answer Quality | No | Yes, with direct answer detection |
| Competitive Analysis | No | Yes, with gap identification |
| Engagement Metrics | No | Yes, with predictions |
- Complete AI Search Audit: Not just schema, but full content optimization
- Content Strategy Planning: Identifies gaps and opportunities
- ROI Prioritization: Know what to fix first for maximum impact
- Competitive Intelligence: Understand your position vs industry standards
- Editorial Guidelines: Use metrics to guide content creation
Same as original - paste into Screaming Frog's Custom JavaScript feature with your Gemini API key.
- Original concept and base code by Metehan Yesilyurt
- check out his orgional blog post here
- Updated by the Hueston LLMO Team @ Robin Adams
08-08-2025 - Tightened FAQ detection: removed '?' from indicators and used regex word-boundary question starters to reduce false positives.