Social listening

In the API this is the signal type social_listening (Social listening). You run it with the same route as every signal: a search.

What it finds

People who post about a topic on LinkedIn, and/or who react to or comment on those posts. Each row is a person whose post matches the words, with the post, without any AI. There is no relevance check, so narrow the search with the query syntax, must_include and exclude, and judge each post yourself.

Rows and price

• One row per person, as the source shows it, with no AI judging it: you decide which rows fit.
• At most 100 rows per search, 25 unless you send limit; more with cursor. 1 credit per row.
• A search that delivers nothing is free.

The signal object

Send it as signal in the body of POST /searches. The other fields of the body are on Start a search.

• type string · required — Always social_listening.
• query string · required — Keywords as the LinkedIn post search reads them, with at most 5 AND, OR and NOT.
• must_include string[] — Words a relevant post also uses; they narrow the search: every post uses one of them. Up to 10 items.
• exclude string[] — Words that only belong to another meaning; posts with them are dropped. Up to 10 items.
• leads_from string — authors = people who wrote a post; reactors = people who reacted to or commented on the busiest posts; both = merged and deduplicated. Defaults to authors. One of: authors, reactors, both.

Limits

• LinkedIn posts of about the last month, as far back as the post search reaches: there is no backfill.
• Every post that matches the words counts, job posts and ads included: there is no relevance check.

Run it

• Once: Start a search (POST /searches). Preview a search shows the most it can cost first. The rows are in List search results.
• To follow it: run the search again, for example every week. There are no watches.
• Keyword monitor: Create a monitor (POST /monitors) reads the LinkedIn posts that use your keywords on a schedule and delivers each new one as a mention, at 1 credit per mention. Read them with List mentions.
• From your assistant: the Signals MCP server (MCP) runs it with search_signals.

Example

A search, POST /searches:

{
  "signal": {
    "type": "social_listening",
    "query": "\"payroll software\" OR \"payroll provider\"",
    "leads_from": "authors"
  },
  "limit": 25
}