LLllmpulse
Prompts

List query fan-out

Changed on

The sub-queries a model actually issued when answering your tracked prompts. view=query (default) returns one row per distinct sub-query with count and share of all occurrences; view=prompt returns one row per prompt with how many distinct sub-queries it produced. Fan-out is reported mainly by ChatGPT, so an empty result usually means the models in scope do not expose it. The API returns the aggregation only: for a period-over-period delta, call it twice with explicit from/to.

get/dimensions/query_fan_outs

Request

  • Base URL
    https://api.llmpulse.ai/api/v1
  • URL
    https://api.llmpulse.ai/api/v1/dimensions/query_fan_outs
  • Auth
    API key in header Authorization

Query parameters

project_idinteger required

Project ID

pageinteger
per_pageinteger
view'query' | 'prompt'

Row shape: one per distinct sub-query, or one per prompt

order'count' | 'share' | 'query_text' | 'total_count' | 'variations_count' | 'prompt_text'

Sort field; the allowed set depends on view

direction'asc' | 'desc'
querystring

Case-insensitive substring filter on the sub-query text

model'chatgpt' | 'perplexity' | 'ai_mode' | 'ai_overview' | 'gemini' | 'copilot' | 'amazon_rufus' | 'claude' | 'grok' | 'deepseek' | 'naver_ai' | 'baidu_ai' | 'meta_ai' | 'naver_ai_tab' | 'qwen' | 'doubao'

Filter by one AI model. It must be a model enabled for the API key's user: an unknown value, or a model the user has not enabled, returns 422 ERR_INVALID_PARAM with the valid values in the message.

collection_idstring

One collection/tag ID or a comma-separated list of IDs. A query value is always a string on the wire, so it is typed as one: the previous integer-or-string union made generators emit a wrapper type they could not serialize.

country_codestring

One ISO country code or a comma-separated list (e.g. US,GB,DE)

language_codestring

One ISO language code or a comma-separated list (e.g. en,es,de)

promptinteger

Filter by prompt ID

prompt_typestring

One prompt type or a comma-separated list: informational, navigational, commercial, transactional

brand_kind'brand' | 'brand_other' | 'non_brand'

Filter by brand kind: brand (own brand/products), brand_other (competitors/other brands), non_brand (generic, no brand named). For fair 1:1 brand-vs-competitor comparisons (visibility, share of voice), use non_brand: brand-focused prompts skew results toward the brand they name. The in-app Overview page applies non_brand by default.

rangeinteger

Number of days to look back (alternative to from/to)

fromstring date-time
tostring date-time

End of the window. A date-only value such as 2026-09-01 covers that whole day. Pass a full timestamp to end the window earlier.

output'flat' | 'csv'

Rectangular output for BI tools (Tableau, Excel, Sheets, ELT). Omit for the default nested JSON. 'flat' returns the same metadata plus 'columns' and 'rows'; 'csv' returns those rows as text/csv. Errors are always returned as JSON.

Response

Paginated fan-out rows

Changes