Lumirank Get the visibility report
/v1/reports/visibilityShows how often AI answers name your brand, how high they rank it, and how positive they are about it. The report has 1 row for each group. A group is the answers that have the same values for the dimensions you ask for, for example the same week and engine. An answer mentions your brand when the brand is in the answer's ranked list, or when its name is in the text. The numbers are the same as the numbers on the Overview page, in the sentiment spread, and on the Prompts page of the dashboard.
Without dimensions, the report has 1 row for the whole date range. You can add up counts across rows, but not rates or averages. For a weekly number, ask for the week dimension instead of averaging 7 day rows.
Authorization
apiKey An API key that you create in the Lumirank dashboard, on the Settings > API keys page. Send it in the Authorization header: Bearer lr_....
In: header
Query Parameters
The first day of the date range, as a UTC date in the format YYYY-MM-DD. The range includes this day. The default is 29 days before end_date, so the default range is 30 days.
dateThe last day of the date range, as a UTC date in the format YYYY-MM-DD. The range includes this day. The default is today. The range can be at most 90 days long.
dateHow to split the report into rows. Send up to 3 of these, separated by commas: day, week, or month (only 1 of these 3), engine, country, and prompt. The report has 1 row for each group of answers that have the same values, for example 1 row for each week and engine. Without it, the report has 1 row for the whole date range.
The engines to include, separated by commas: chatgpt, gemini, perplexity, ai_mode, copilot, google_ai_overview, claude. Without it, every engine is included.
The countries to include, separated by commas, as 2-letter ISO 3166-1 country codes in capital or small letters. Without it, every country is included.
The ids of the prompts to include, separated by commas. Without it, every prompt that is not archived is included. The id of an archived prompt gets status 400 unless you send include_archived=true.
The intents of the prompts to include, separated by commas: brand, comparison, how_to, informational, other. The intent of a prompt is what its question asks for. Intent is only a filter: you cannot use it as a dimension.
Send true to include the answers of archived prompts. The default is false: they are left out, as in the dashboard.
falseHow to sort the rows. Send up to 3 columns, separated by commas. Add :asc after a column for the smallest values first, or :desc for the largest first. Without either, the order is asc. You can sort by any column that the report returns. The default order is by the time dimension if you ask for one, then by answers, largest first.
The number of rows on each page, from 1 to 10000. The default is 1000.
int32The number of rows to skip before this page starts, from 0 to 100000. The default is 0.
int32Response Body
application/json
application/json
application/json
application/json
application/json
application/json
application/json
curl -X GET "https://example.com/v1/reports/visibility?start_date=2026-09-01&end_date=2026-09-30&dimensions=week%2Cengine&engine=chatgpt%2Cperplexity&country=US&prompt_id=a41e7c52-3d6b-4f0a-9c8e-1b2d3e4f5a60&intent=comparison&include_archived=true&order_by=visibility%3Adesc&limit=1000&offset=0"weekly by engine
{ "data": [ { "week": "2026-09-07", "engine": "chatgpt", "answers": 412, "mentions": 173, "visibility": 0.4199, "visibility_index": 0.3012, "average_position": 2.4, "best_position": 1, "sentiment": 38.5, "sentiment_positive": 120, "sentiment_neutral": 40, "sentiment_negative": 6, "own_citations": 57 } ], "meta": { "report": "visibility", "start_date": "2026-09-01", "end_date": "2026-09-30", "dimensions": [ "week", "engine" ], "engine": [ "chatgpt" ], "include_archived": false, "order_by": [ "week:asc", "answers:desc" ], "limit": 1000, "offset": 0, "total": 5, "has_more": false, "data_through": "2026-09-30T06:41:12Z" }}Get the citations report GET
Shows the websites and pages that AI engines cite for your prompts. The report has 1 row for each cited domain, or 1 row for each cited page with the `url` dimension. The counts are the same as in the domain and URL tables on the Citations page of the dashboard. If `dimensions` has neither `domain` nor `url`, the report adds `domain`. So `dimensions=week` returns 1 row for each week and domain. In each group, the `citation_share` of all rows adds up to 1. `citation_rate` divides by the `answers` that the visibility report gives for the same group. The citations in the answers of archived prompts are left out unless you send `include_archived=true`. This applies to the citations and to the answers that `citation_rate` divides by.
Changelog
Changes to the Lumirank API, newest first.