Design a page
AI engines can cite.

What Citation Studio is

Every other IndexHalo feature audits content that already exists and recommends fixes. Citation Studio inverts the pipeline: you compose a web page inside the tool, and the full measurement engine — signal scoring, RAG chunk simulation, retrieval drift, schema compilation — runs live at authoring time as a co-writer. The page is born citation-ready instead of retrofitted.

The mental model is simple: a page is an ordered list of measurement-backed blocks. Each block type is a citation pattern the engine already knows scores well — answer capsules, FAQ schema, sourced figures. You fill in facts; the compiler renders semantic HTML and JSON-LD; the scorer replays the same analysis engine against the compiled output and reports per-block verdicts on every edit.

The four free block types

The free preview covers the four block types that contribute most to citation readiness. Each one maps to a signal the engine measures:

Definition head. A 40–80 word self-contained opening passage that names the subject and audience. Feeds retrievability and specificity — the signals that determine whether an engine can extract a clean answer from the top of your page.

Answer capsule. A question paired with a 40–80 word answer that needs no surrounding context to make sense. This is the pattern engines lift verbatim — it feeds retrievability directly.

FAQ. Up to three question-answer pairs that mirror real user queries, rendered as FAQPage schema. Feeds structure and retrievability.

Stat callout. A single standout figure with context, source, and date. Feeds specificity — dated, sourced figures are the strongest signal that your content is worth quoting.

How the score works

When you press Score this page, the tool compiles your blocks into semantic HTML and a JSON-LD graph, then runs the same analyzeMetadata engine that powers full IndexHalo reports. You get six measured signals — retrievability, authority, specificity, structure, freshness, transparency — plus a chunking analysis that shows whether each passage fits inside a model's context window, and a retrieval-drift check against your target queries.

Every score carries its evidence class. Signal scores and chunking are measured — deterministic over the compiled text. Retrieval drift is measured when an embedding provider is configured and modelled otherwise, labelled as such. The free preview does not include the live entity probe (observed) or FactBase consensus — those need a brand profile and a saved page, which is what the paid Studio provides.

What the full Studio adds

The free tool is the same engine, capped at four block types and two queries, with no persistence. The dashboard Studio unlocks all fourteen block types, saved pages (5 on Professional, 50 on Agency, unlimited on Enterprise), a live entity audit that probes whether engines can reconcile your organisation, FactBase consistency checks that flag claims contradicting your brand's recorded facts, the full six-file export bundle, and publish verification — a parity check that compares your live page to the scored version and creates a before/after experiment. See plans and limits →