recent-tweet-search is for a question that comes back. Save a query for a brand, topic, or phrase, then repeat it with the same query, language, and max_results values. That gives you two result sets you can compare.

The route returns recent public post text, timestamps, public metrics, and expanded author profiles. It stays read-only. It never posts, follows, likes, or reads direct messages.

If native X search can't answer, the response keeps its tweets array and sets degraded: true. Those public-index results can still be useful for a lead, but they aren't a like-for-like substitute when you need native X metrics or full author expansion.

The request fields, response shape, and price are unchanged. This is clearer catalog and MCP copy for agents returning to the same monitoring question.