GeoHarness

GEO score methodology and the Lighthouse gate

GeoHarness scores how well generative engines can reach, index and read a site. Most of the score rests on access and delivery checks backed by platform documentation; content heuristics are advisory because the research does not show they reliably move citations.

Updated · Every source is on the citations page with its evidence grade.

Why are the weights what they are?

A survey of 45 GEO studies finds no content technique with a stable cross-platform effect; relevance and position are the reproducible levers, and a page that cannot be fetched, indexed or extracted has neither. Rewrite-style tactics mostly fail under benchmark conditions. So 80 of 100 points sit on access, indexability and markdown delivery, and those checks are blocking: a failure stops the run whatever the score. The remaining 20 points are advisory.

GEO score categories
CategoryIdWeight
AI accessaccess30
Indexabilityindexability25
Markdown deliverydelivery25
Extractability (advisory)extractability12
Optional signals (advisory)signals8

Which checks run, and what evidence backs each?

Blocking checks carry grade A or B evidence. Checks with live in their id run after deploy against the real site, with the Accept matrix and eight AI user agents.

Every check, its severity and the evidence behind it
CheckWhat it verifiesCategorySeverityEvidence gradeCitations
access.spa_shellPage has server-rendered textaccessblockingA, Cpeec-how-deep-research-reads-2026, google-ai-guide-2026
access.robots_botsrobots.txt admits search and user agentsaccessblockingAopenai-bots, anthropic-bots, perplexity-bots, google-ai-guide-2026
access.live_ai_agentsAI user agents get HTTP 200, not a challengeaccessblockingAcf-bot-fight-mode, openai-bots, anthropic-bots, perplexity-bots
index.titlePage has a <title>indexabilityblockingAgoogle-ai-guide-2026
index.descriptionPage has a meta descriptionindexabilityblockingAgoogle-ai-guide-2026
index.canonicalPage has a canonical linkindexabilityblockingAgoogle-ai-guide-2026
index.broken_linksInternal links resolveindexabilityblockingAgoogle-ai-guide-2026
index.live_buildLive site serves this buildindexabilityblockingAcf-workers-static-assets
delivery.twinMarkdown twin exists and matches the pagedeliveryblockingA, A (spec)google-spam-policies, llmstxt-spec, htmlrag-2025
delivery.live_negotiationAccept: text/markdown gets markdown with Vary: AcceptdeliveryblockingA, Ccf-markdown-for-agents, cf-transform-rules, checkly-2026
delivery.live_filesHTML pages, robots.txt, llms.txt and sitemap.xml are served correctlydeliveryblockingA (spec)cf-content-signals, llmstxt-spec
extract.link_countFew links (<20) so the first read covers the pageextractabilityadvisoryCpeec-reading-budget-2026
extract.front_loadedMain content starts within the first ~5,700 charactersextractabilityadvisoryCpeec-reading-budget-2026
extract.entity_earlyEntity named in the first 30% of the contentextractabilityadvisoryCindig-2026
extract.question_headingsA heading is phrased as a questionextractabilityadvisoryA, B (preprint)what-gets-cited-2026, sageo-2026
extract.visible_dateA date is visible in the contentextractabilityadvisoryB (preprint)what-gets-cited-2026
extract.tables_keptTables survive into the twinextractabilityadvisoryAhtmlrag-2025
signals.llms_txtllms.txt is present and well-formedsignalsadvisoryA (spec), Bllmstxt-spec, ahrefs-llmstxt-2026, seranking-llmstxt-2026
signals.json_ldJSON-LD is present and validsignalsadvisoryA, Bahrefs-schema-2026, google-ai-guide-2026

How is the score computed?

Each finding scores 0 to 1; a category's score is the mean of its findings, and the GEO score is the weighted mean over the categories that had findings. The build audit runs on the processed output; the live audit adds the post-deploy observations. geo-min-score sets the minimum for both.

What does the Lighthouse gate require?