What GET /2/tweets/search/recent actually costs under X pay-per-use credits, and when a ranked X trending feed is cheaper than reconstructing trends from tweet reads.
Live data as of 2026-07-30
X API recent search (GET /2/tweets/search/recent) is billed under pay-per-use credits as of 2026, with third-party post reads listed around $0.005 each. A single request that returns 100 posts costs about $0.50 before expansions. The endpoint only covers roughly the last 7 days. For teams that need a ranked list of what is trending on X rather than the tweets themselves, Trends MCP returns 25 live topics from type X (Twitter) Trending in one call, and the free plan covers 100 requests per month.
From the Trends MCP scan of July 30, 2026 (as_of_ts 2026-07-30T06:01:37Z), the top 15 topics on X (Twitter) Trending were:
| Rank | Topic |
|---|---|
| 1 | #AEWDynamite |
| 2 | Ali Sanchez |
| 3 | Romy |
| 4 | Boone |
| 5 | Thornton |
| 6 | O'Hoppe |
| 7 | Tim Hill |
| 8 | Doval |
| 9 | Edwin Diaz |
| 10 | Arike |
| 11 | Silseth |
| 12 | Edgar Quero |
| 13 | Jordan Hicks |
| 14 | Eric Lauer |
| 15 | #WhiteSox |
Ranks 16 through 25 completed the board with #brandnewday, Andruw Monasterio, Ospreay, Angels, Valkyries, #WWHL, Grisham, Allisha Gray, Teoscar, and Rhyne Howard. That is 25 distinct topics from one get_top_trends call, with no per-tweet credit burn on the X developer meter.
Google Search growth follow-ups on the same day (Trends MCP API, July 30, 2026) show how board presence and search demand can diverge. AEW Dynamite fell 30.61% over 7 days (index 49 to 34) and 35.85% over 30 days. White Sox rose 38.71% over 7 days (31 to 43) while slipping 6.52% over 30 days. Spider-Man Brand New Day jumped 77.78% over 7 days, 113.33% over 30 days, and 540% over 3 months (index 5 to 32).
Recent search charges scale with returned posts. At $0.005 per third-party post read, 10 results cost $0.05, 50 results cost $0.25, and the 100-result maximum costs $0.50 per successful call. Owned reads can list lower (about $0.001 in some console tables), but monitoring public conversation usually hits the third-party rate. Deduplication rules may waive a second charge for the same resource inside a 24-hour UTC window; verify that behavior in the console for the current app.
A practical monitoring loop makes the math concrete. Pulling 50 queries once per hour at 100 posts each is 1,200 searches and up to 120,000 post reads per day. At $0.005 that is $600 per day, or about $18,000 per 30-day month, before user profile lookups at about $0.010 each. Even a lighter plan of 20 queries every 6 hours at 50 posts each is 2,400 searches and 120,000 reads per 30 days, still about $600 per month in post-read credits alone.
Recent search is not full archive. Queries only see roughly the prior 7 days, so long-horizon brand studies need a different endpoint or vendor. Rate limits sit in 15-minute rolling windows. Planning docs often cite about 450 recent-search requests per 15 minutes per app, with a default page size near 10 and a maximum of 100. Those ceilings cap throughput, but they do not cap the credit bill: hitting 450 requests at 100 posts each can still consume 45,000 post reads in a quarter hour, about $225 at $0.005.
Teams building "what is trending" detectors on top of recent search also pay for false positives. Every exploratory query that returns noise still burns credits. The Trends MCP board already ranks topics; a follow-up search is optional when the product only needs the topic names.
One get_top_trends call returns the live board. Growth on related search demand uses a separate source such as google search, because X trending on Trends MCP is a live feed, not a historical keyword series.
{
"mode": "get_top_trends",
"type": "X (Twitter) Trending",
"limit": 25
}
{
"mode": "get_growth",
"source": "google search",
"keyword": "Spider-Man Brand New Day",
"percent_growth": ["7D", "30D", "3M"]
}
On July 30, 2026, the first call returned 25 ranked topics. The second returned three growth windows: +77.78% (7D), +113.33% (30D), and +540% (3M). Pairing the board with growth separates a one-day spike from a multi-week climb without downloading thousands of tweets.
Recent search remains correct when the product needs tweet text, author IDs, conversation threads, or engagement fields. Social listening that classifies sentiment on individual posts cannot replace those fields with a trending-topics list. Posting bots, CRM sync, and compliance exports also need the X API directly.
The cut is narrow. If the question is "which 25 topics are trending on X right now," paying per post is the expensive path. If the question is "what did these 500 accounts say about brand X in the last 48 hours," recent search (or a stream) is the fit, and the credit budget should be modeled before launch. For plan-level context across all X endpoints, see the X API pricing comparison. For the Trends MCP feed shape, see the X trending topics API.
Trends MCP returns ranked topic strings, not tweet IDs or engagement counts. The July 30, 2026 board was heavy on live sports and entertainment names; composition changes by hour. Google Search growth is a correlated signal, not X native volume. X credit prices and rate limits change; any dollar figure on this page should be re-checked in the Developer Console before a procurement decision. Trends MCP pricing for comparison: Free at 100 requests per month, Starter at $19 for 1,000, Pro at $49 for 5,000, Business at $199 for 25,000 (pricing). Related signal pages include Twitter / X trends.
FAQ