How ConduitScore Is Calculated
The ConduitScore 0–100 score measures how well any website can be discovered, read, and cited by AI agents. It is based on 7 core categories that most directly affect AI visibility — each weighted by its real-world impact on AI discoverability.
What Your Score Means
Core Score Methodology
The ConduitScore is based on 7 core categories. Each category contributes a maximum number of points toward the 100-point total. A category score reflects how well a site satisfies the checks within that category — from fully passing (full points) to completely absent (zero points).
Crawler Access
Checks whether AI bots (GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot) can access your site — including not just the absence of blocks, but the presence of explicit Allow rules that signal proactive AI-friendliness. Also checks robots.txt configuration and whether a sitemap is discoverable. LLMs.txt vs robots.txt guide →
Structured Data
Evaluates JSON-LD schema markup including Organization, WebSite, FAQPage, BreadcrumbList, and article schemas that help AI systems understand your content. Schema for AI citations guide →
LLMs.txt
Checks for the presence, completeness, and structure of your /llms.txt file — a machine-readable guide for AI agents about your site's key pages and purpose. Also checks for a /llms-full.txt companion file and HTML meta tags (link rel="llms-full" and link rel="agent-manifest") that help autonomous AI agents discover your site's capabilities. LLMs.txt implementation guide →
Content Structure
Assesses heading hierarchy, semantic HTML, FAQ sections, introductory paragraphs, and answer-friendly formatting.
Technical Health
Evaluates page load speed, canonical tags, viewport meta, meta descriptions, noindex directives, and other technical signals affecting AI crawler success.
Citation Signals
Checks for author attribution, external links to authoritative sources, about/contact pages, trust/legal pages, and clear organization identity.
Content Quality
Measures content depth, title and description quality, publish dates, paragraph structure, and extractability for AI answer generation.
Score Range Interpretation
Use the following ranges to interpret any ConduitScore result:
Additional Diagnostic Modules
Some scan reports include additional diagnostic checks beyond the 7 core categories. These modules provide deeper insight into specific areas but do not currently change the main 0–100 score. They are provided as supplemental guidance to help you identify advanced optimization opportunities.
AI Bot Policy
A detailed breakdown of which AI crawlers are allowed or blocked, including OAI-SearchBot, GPTBot, and others.
Answer Extraction Readiness
How easily an AI system can extract a reliable answer from your content — based on heading quality, lists, tables, and paragraph structure.
Public Reportability Gap
Whether your site has the key pages (methodology, examples, about) that make your business easy for AI systems to summarize and recommend.
Learn More
Frequently Asked Questions
How is the ConduitScore calculated?
ConduitScore checks 46 signals across 7 categories: Crawler Access (15 pts), Structured Data (20 pts), LLMs.txt (10 pts), Content Structure (15 pts), Technical Health (15 pts), Citation Signals (15 pts), and Content Quality (10 pts). Each category is scored 0–100 independently, then weighted by its point allocation to produce the overall 0–100 score.
What counts as a good AI visibility score?
Scores of 75–89 are considered Good — well-optimized with minor improvements available. Scores of 90–100 are Excellent. The median score across all scanned sites is around 28/100, so even a score of 50 puts you well above average.
Does a high ConduitScore guarantee my site will appear in AI answers?
No. AI visibility depends on many factors including content quality, domain authority, and whether your topic appears in the training data and retrieval indexes of each AI system. ConduitScore measures the technical and structural signals that make it possible for AI to access, read, and cite your content — but it cannot guarantee specific AI system behavior.
What is llms.txt and why does it matter?
llms.txt is a plain-text file placed at the root of your website (yoursite.com/llms.txt) that gives AI agents a structured guide to your site's key pages, purpose, and preferred sections. It is similar to robots.txt but designed for LLMs rather than search crawlers. Sites with a well-structured llms.txt make it significantly easier for AI systems to index and cite them accurately.
How do I improve my score?
The highest-impact fixes are typically: (1) adding or correcting structured data (JSON-LD schema markup), (2) creating an llms.txt file, (3) fixing robots.txt to explicitly allow major AI bots (GPTBot, ClaudeBot, PerplexityBot), and (4) adding clear author attribution and contact information. Each scan returns prioritized, copy-paste fixes for your specific site.
How often should I scan my site?
We recommend scanning after any major site change — redesign, new content sections, robots.txt updates, or structured data changes. Paid plans include weekly automated scans and score-drop alerts.
What signals does ConduitScore check for Crawler Access?
The Crawler Access category (15 pts) checks whether major AI bots — GPTBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot, OAI-SearchBot, and GoogleBot — are explicitly allowed in your robots.txt. It also verifies that a sitemap.xml exists and is referenced from robots.txt, and checks for explicit Allow directives rather than just the absence of a Disallow rule. Proactive Allow rules signal AI-friendliness beyond passive permission.
What is the Open Knowledge Format (OKF) and why does it matter?
OKF (Open Knowledge Format) is a structured bundle of AI-readable markdown files placed at /okf/ on your website — introduced as a Google standard in June 2026. The bundle includes files like /okf/index.md, /okf/methodology.md, /okf/faq.md, and /okf/pricing.md that give AI agents a machine-readable summary of your business. Sites with an OKF bundle score higher on the LLMs.txt category and are more likely to be cited accurately by AI systems.