Citino API

The Citino API gives your scripts, warehouse jobs and spreadsheets the same data the dashboard shows. Every figure comes from the same calculation the screen uses, so a number you pull matches the number on Overview or Brands for the same scope.

​

What you can do

• Read every screen’s data. Your headline figures, the league table, the daily trend, visibility by engine and by segment, what moved, the prompts you are losing, Prompt detail, Sources and Source detail, engine searches, Objections, Crawlability, Crawler activity, Attribution, Alerts, Experiments, Gaps, Claims, and an agency’s monthly reports.
• Pull report rows. Visibility, share of voice, position and sentiment for every brand named, and retrieval and citation rates for every domain and URL, broken down by date, engine, segment and prompt. See Reports.
• Read the answers themselves. List the answers collected for a project, and fetch one with its full text.
• Keep your setup in step from code. Add, re-word and archive prompts; track and stop tracking competitors; add and remove brand aliases; set the segments you claim and the engines you sample. See Writes.
• Download CSV. The list and report routes answer format=csv, so a page drops straight into a spreadsheet. See CSV output.

​

The basics

Requests and responses are JSON (or CSV where you ask for it). Field names are snake_case, with one exception: a scope object that describes what a screen’s figures cover is passed on as the screen states it, so treat its keys as information rather than a fixed shape. Dates are YYYY-MM-DD and are days in the project’s market time zone.

​

What the numbers cover

Report rows and the headline figures count your unbranded prompts: the questions that do not name your brand or a competitor you track. Branded and comparison questions are listed by the prompts route, kept apart, and never blended into the headline figures. Only the segments you have claimed count towards your scores unless you ask for others. The headline figures and the report rows come with a scope object saying which days, engines and segments they counted.

​

Answer text is somebody else’s writing

One route returns an engine’s answer in full: fetching a single answer with include=answer. That text was written by an AI engine from pages anyone can publish. Treat it as untrusted text: store it and display it as text, and never execute it or pass it to a system that follows instructions. No other route, and no MCP tool, returns answer text.

​

Where to go next