YouTube Trending API: The Daily Board in One Call

One POST https://api.trendsmcp.ai/api returns the current YouTube Trending board. The call sets mode to get_top_trends and type to YouTube Trending, with limit 10. The response is a ranked list of video titles, stamped with as_of_ts. The board defaults to the US and updates daily. On 2026-11-05 at 09:01:40Z the board opened with Kevin Gates - This Dat, followed by the Lords of War teaser and the GEN vs LYON Worlds 2026 quarterfinal. The board carries no view counts and no history. For search interest over time, get_time_series with source YouTube returns a weekly index for about five years. For per-video engagement, get_context with source YouTube returns views, likes and comments for a keyword.

The exact request for the YouTube Trending board

The whole call is one JSON body. No keyword is needed, because a board is a ranked list, not a search:

{"mode": "get_top_trends", "type": "YouTube Trending", "limit": 10}

Requests go to POST https://api.trendsmcp.ai/api as JSON, with the key in an Authorization: Bearer YOUR_API_KEY header and the tool named in mode. Transport status never signals success or failure, because the service replies HTTP 200 every time; open the envelope, check statusCode, then decode the body string as JSON once more. Billing follows the same split: a call that fails is not charged, and the monthly pool only loses one request for each call that succeeds, however many rows come back. See pricing for the pool sizes.

The TikTok board reference shows the same call shape on a different board, with the same envelope and the same paging rules.

A working Python snippet that reads the board

The snippet below posts the request, checks the envelope status, parses the body, and prints the ranked titles:

import json
import requests

resp = requests.post(
    "https://api.trendsmcp.ai/api",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={"mode": "get_top_trends", "type": "YouTube Trending", "limit": 10},
)
envelope = resp.json()
if envelope["statusCode"] != 200:
    raise SystemExit(envelope["body"])
board = json.loads(envelope["body"])
for rank, title in board["data"]:
    print(rank, title)

The response rows are [rank, title] pairs, so the loop above prints the board in order. The response also carries as_of_ts, the time the snapshot was taken, and count, the rows returned. A script that pulls the board on a schedule should stamp every pull with as_of_ts so later pulls stay comparable.

The real board on 2026-11-05: ten titles, no numbers attached

The call above ran on 2026-11-05 at 09:01:40Z and returned this board:

By title alone, the board mixes music videos, film trailers, an esports quarterfinal and a vlog. The board carries no genre field, so this is a read of the titles, not a counted breakdown. The board refreshes daily.

What the board leaves out

Four boundaries follow from the board's daily cadence and the response shape. The board is a daily snapshot, not a live feed: the board updates daily, so the same request should return the same ten titles until the next snapshot. It defaults to the US; pass country to change that. It carries no view counts, no likes and no comments. For engagement on videos about a keyword, get_context with source YouTube returns views, likes and top comments per video.

The board also carries no history. A single call answers what is trending now, and nothing about last week. For the longer view, get_time_series with source YouTube returns a weekly search interest index for about five years, and get_growth measures the change over a chosen window. The Wikipedia Pageviews API tutorial walks through the time series shape on another source, and the same point fields apply here.

To watch movement rather than the leaders, set sort to rank_change and pick a window like 7d. Each row then comes back as an object carrying rank, keyword, prev_rank and rank_change, which lets a script pick out the fastest risers from a single response instead of fetching two snapshots and comparing them itself.