Mentionable
Mentionable

list_fan_outs

list_fan_outs MCP tool: deduplicated fan-out queries the LLMs ran behind your tracked prompts, ranked by frequency, with first/last seen dates and the LLMs involved.

Updated 2026-04-26

list_fan_outs

list_fan_outs exposes the raw search queries that LLMs run internally when they answer your tracked prompts. Each row is a deduplicated query, ranked by frequency, with the LLMs that issued it, the prompts that triggered it and first/last seen dates.

When to use

Use this tool when the user wants to understand "what is the LLM actually searching for on my behalf?". It is the best tool for content discovery: every recurring fan-out query is a content angle the LLMs already pull external sources from. Combine with list_llm_sources to see who is winning those searches today, and with list_prompts to trace a query back to the user-facing prompt that triggered it.

Input

Field Type Default Description
projectId string (CUID) required Project to query.
cursor string — Numeric offset cursor.
limit integer 20 1 to 100.
filters.search string — Substring match on the query text.
filters.promptId string (CUID) — Restrict to fan-outs for one prompt.
filters.llm enum — Restrict to one LLM.
sortBy enum frequency frequency or recent.

Queries shorter than 4 characters are dropped server-side.

Response

data is an array of deduplicated query objects. occurrences is the number of times the query was issued across all tracked prompts and LLMs.

{
  "data": [
    {
      "query": "ai visibility tracking software",
      "occurrences": 23,
      "llms": ["CHATGPT", "GEMINI", "PERPLEXITY"],
      "promptIds": ["clxprompt001", "clxprompt007"],
      "firstSeenAt": "2026-03-15T06:00:00.000Z",
      "lastSeenAt": "2026-04-25T06:00:00.000Z"
    }
  ],
  "pageInfo": { "hasMore": true, "nextCursor": "20", "totalCount": 142 }
}

Tips and patterns

  • Sort by frequency to surface the queries the LLMs run most often. These are the highest-leverage content topics.
  • Filter by llm to compare what each engine searches for the same prompt set. Differences expose engine-specific blind spots.
  • Use filters.promptId to drill down from a single prompt to the fan-outs that drive its answer.
  • A query with many llms and a recent lastSeenAt is a stable, cross-engine intent signal worth covering with dedicated content.

Related tools