Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.getcited.in/llms.txt

Use this file to discover all available pages before exploring further.

Defining what a product does not do is as important as defining what it does. This page documents the deliberate scope boundaries Cited has drawn — things intentionally excluded from measurement, features not built, and claims not made. Each boundary has a specific reason.

​
Google AI Overviews are not actively tracked

AI Overviews (AIO) were tested via SerpAPI in the India market. Result: approximately 3% of tracked queries triggered an AI Overview — too sparse for meaningful benchmarking. The AIO data from historical runs is preserved but excluded from active monitoring metrics. This may change as Google expands AIO coverage across more query types and markets. The infrastructure to resume AIO tracking exists; the decision not to run it is a signal-quality judgment, not a technical limitation.

​
Deterministic measurement is not claimed

LLM responses are inherently non-deterministic. Cited does not claim to provide exact, repeatable measurements. Instead, Cited provides aggregated metrics that are statistically stable at sufficient sample sizes. This is an honest limitation of the measurement domain, not a fixable bug. The same principle applies to any tool measuring AI-generated content — the underlying systems produce variance by design.

​
Brand content is not graded

Cited measures brand visibility in AI answers. It does not grade, score, or evaluate the quality of a brand’s own content (blog posts, product pages, documentation). Content quality assessment tools exist; Cited is not one of them. The GEO Score product provides a technical readiness assessment of a brand’s web presence for AI discoverability — checking robots.txt configuration, schema markup, llms.txt presence, and HTML structure. This is a technical audit, distinct from content quality grading.

​
Visibility improvements are not guaranteed

Cited provides measurement and diagnostics. It does not guarantee that following recommendations will produce a specific increase in mention rate or share of voice. AI platform behavior is controlled by the platforms themselves — OpenAI, Google, Anthropic, Perplexity, xAI. No external measurement tool can guarantee outcomes in a system it does not control. Cited identifies where opportunities exist and measures whether changes are working, but the outcome depends on the AI platforms’ responses to content and authority signals.

​
Impact Scores are brand-relative, not predictive

Impact Scores are brand-relative and do not predict exact visibility lift. A score of 60 means “60% of the maximum achievable impact for this brand’s current data maturity” — it does not guarantee a specific mention rate or share of voice increase. The estimated_improvement field on each recommendation provides a separate, rougher directional estimate.

​
Benchmark data is not fabricated

Every benchmark number published in Cited’s documentation comes from the Cited Index pipeline — real queries run against real AI platforms, parsed and aggregated from real responses. No number is estimated, projected, or synthesized from secondary sources. The benchmarks methodology page describes exactly how the data is produced, including sample sizes, query counts, and caveats about statistical precision. When data quality is insufficient for benchmark publication (as with sentiment and citation rate in the current edition), this is documented transparently rather than filled with estimates.

​
Individual user conversations are not tracked

Cited queries AI platforms via API and measures the responses. It does not monitor, intercept, or analyze individual user conversations with AI platforms. The data is about brand presence in AI answers, not about user behavior or user identity.

​
Custom JSON-LD schema is not currently implemented

Due to a platform limitation with Mintlify’s React Server Components pipeline, Cited’s documentation does not inject custom JSON-LD schema markup. This is documented in detail on the schema platform tradeoffs page. Cited’s docs rely on auto-generated llms.txt and llms-full.txt, textual citability patterns, clean HTML structure, and a comprehensive crawler allowlist for LLM discoverability. These mechanisms are working effectively — the schema limitation is a known tradeoff, not an oversight.

​
Frequently asked questions

Likely yes, as Google expands AIO coverage beyond the current trigger rate in the India market. When AIO produces enough data to support meaningful benchmarking, Cited will integrate it. The historical AIO data from earlier testing has been preserved for this purpose.
Because measurement credibility depends on honesty. A tool that claims to measure everything perfectly is not trustworthy. Documenting scope boundaries explicitly helps users interpret Cited’s data correctly and make informed decisions about what the data can and cannot tell them.
Yes. The development roadmap prioritizes features by customer demand and data viability. Planned additions include Grok integration into the Cited Index, expanded query libraries, and citation source mapping with PR target recommendations. These are prioritized against customer value and engineering cost.