qalarc.com / projects / universal-geo-analyzer
Universal GEO Analyzer
Universal GEO Analyzer — by Qalarc — measures what AI engines actually say about a company: generates strategic test queries, asks GLM/OpenAI/Anthropic-class engines, and scores mention rate, accuracy and sentiment. Live baseline (Sep 2026): 0% Qalarc mention rate across 6/6 GLM responses — the number that started the GEO program.
Universal GEO Analyzer — the measuring half of Qalarc's generative-engine optimization system (the producing half is the GEO Pipeline). Traditional SEO asks "how do we rank in search results?" — GEO asks "what do AI engines SAY when someone asks about us?", and that is what this tool measures.
How it works. Give it a company name and description. It generates strategic test queries across categories (business, technical, competitive, use-case), sends them to multiple AI engines (GLM, OpenAI, Anthropic-class APIs; a mock mode enables offline UI and test runs via a GLM keychain), then analyses every response: is the company mentioned at all, is the description accurate, is the sentiment right, which competitor gets named instead, and which sources does the engine cite.
Real measurements, not demos. The September 2026 baseline run captured Qalarc's starting position: a 0% mention rate across 6/6 GLM responses — plus live evidence of why GEO matters at all: Brave's AI answer engine mis-attributed the Qalarc Token Killer to a stranger. The web dashboard tracks mention rate over time (history + baseline panels), so every surface fix can be verified as a moving number, not a feeling.
Associated system. It closes the loop with the GEO Pipeline (which generates llms.txt, sitemaps, JSON-LD and static pages that the engines read) and the SEO_strategy program that operates both. Pipeline produces → Analyzer measures → findings drive the next fix.
The GEO system: measure → fix → re-measure
This tool is the measuring half of Qalarc's generative-engine optimization system. The producing half is the GEO Pipeline (llms.txt, sitemaps, 86 static JSON-LD pages, AI-crawler robots rules — all live on qalarc.com). Together with the SEO_strategy program that operates both, the system runs a loop: Pipeline produces → Analyzer measures → findings drive the next fix.
The founding measurement (September 2026): 0% Qalarc mention rate across 6/6 GLM responses — and a live attribution failure, where Brave's AI engine attributed the Qalarc Token Killer to a stranger because the underlying pages were invisible to non-JS crawlers. Both were fixed by the pipeline's static-page wave; the analyzer now guards against regressions by re-measuring after every change.
See it live: qalarc.com/llms.txt · GEO Pipeline.
Capabilities
Universal input: analyses ANY company, not a fixed brand
Intelligent query generation across business / technical / competitive / use-case categories
Multi-AI testing: GLM (production), OpenAI / Anthropic-class APIs, mock mode for offline
GLM keychain: multiple keys with rotation; live and mock status shown per response
Mention-frequency analysis: does the engine name you at all
Accuracy scoring: is what the engine says about you correct
Sentiment analysis of AI-generated descriptions
Competitive analysis: who gets mentioned instead of you
Source attribution: which pages the engines cite as evidence
Web dashboard with history and baseline endpoints — mention rate tracked over time
Live baseline captured Sep 2026: 0% Qalarc mention rate (6/6 GLM responses) — the founding measurement of the GEO program
Export + reporting for shareable GEO audits
Tags
Status: active · Last updated: 2026-09-24