Multi-source growth API in one MCP call

A multi-source growth call on Trends MCP passes a comma-separated source string to get_growth and returns per-source windows plus a trend_analysis block. On the August 24, 2026 pull, chatgpt across google search, youtube, wikipedia returned sources_successful 3, uptrend_ratio 3/3, and wikipedia page-view volume 6,574 on 2026-07-01 (index 47.5, up 41.79% from 33.5 on 2026-06-01). Those three 30D windows do not share a recent_date: wikipedia used 2026-07-01, while google search and youtube used 2026-08-22. stanley cup on amazon, tiktok, youtube processed 3 sources and marked tiktok unavailable. The free plan includes 100 requests per month.

What multi-source growth data shows right now

From the Trends MCP pull of August 24, 2026, comma-separated get_growth calls returned per-source indexes, some absolute volumes, and a top-level trend_analysis object that counts uptrends across sources whose 30D windows do not share a recent_date, which is why chatgpt can print uptrend_ratio 3/3 while wikipedia is still on 2026-07-01 and search is on 2026-08-22.

Per-source 30D prints from the same pull:

wikipedia chatgpt used 44 data points. steam Elden Ring used 53. google search, youtube, and google shopping used 261. amazon used 49. A 3/3 uptrend_ratio on chatgpt is three rising 30D prints on three clocks, not one shared month. For search-only windows, see the Google Trends API.

Exact multi-source get_growth calls

One get_growth body can list several sources in a comma-separated source string, and the response sets metadata.mode to multi_sources, lists successful_sources and failed_sources, and keeps each source's own recent_date inside source_results so a 30D print on wikipedia is not silently aligned to search.

{
  "mode": "get_growth",
  "source": "google search, youtube, wikipedia",
  "keyword": "chatgpt",
  "percent_growth": ["30D", "3M", "6M"]
}

{
  "mode": "get_growth",
  "source": "amazon, tiktok, youtube",
  "keyword": "stanley cup",
  "percent_growth": ["30D", "3M", "6M"]
}

{
  "mode": "get_growth",
  "source": "steam, youtube, google search",
  "keyword": "Elden Ring",
  "percent_growth": ["30D", "3M", "6M", "12M"]
}

Keyword rules still apply per source. steam wants a game display name. wikipedia wants a short article title. amazon wants a product token. Mixing com.openai.chatgpt into a google search slot will miss. When one source in the list is unavailable, the others still return, as tiktok did for stanley cup with status unavailable.

Why uptrend_ratio can lie across calendars

trend_analysis.uptrend_count scores each source's first listed window, and those windows do not share recent_date, so a 3/3 uptrend_ratio on this pull can mix a wikipedia month that closed on 2026-07-01 with a google search month that closed later on 2026-08-22. chatgpt looked like a clean 3/3 rise: google search +1.28% to 79.0 on 2026-08-22, youtube +67.86% to 47.0 on 2026-08-22, wikipedia +41.79% to 47.5 on 2026-07-01. wikipedia's 30D is June 1 to July 1. Search and YouTube's 30D is July 25 to August 22. Publishing "chatgpt is up on every channel this month" from that block drops the date split.

crocs is the opposite lesson. uptrend_ratio 2/3 because Shopping +30.0% (52.0 from 40.0) and YouTube +4.71% (89.0 from 85.0) sit next to Amazon -27.95% (9,187,782 from 12,765,263). volume_weighted_growth printed -28.03 with weight_coverage 1/3, because only amazon returned volume. The weighted figure is the Amazon print, not a blend of three channels. nike uptrend_ratio 1/2 pairs google search +8.47% (64.0 from 59.0) with youtube -4.55% (42.0 from 44.0). google search volume 27,500,000 on that row is volume_estimated true, with volume_growth omitted because volume is derived from the trend value. Do not treat 27,500,000 as an independent query count.

Elden Ring uptrend_ratio 0/3: steam index 1.9 / 46,694 peak players on 2026-07-01 vs 2.4 / 49,524 on 2026-06-01, youtube flat at 6.0 on 2026-08-15 vs 2026-07-18, google search 4.0 vs 5.0 (-20.0%) on the same August dates. steam 12M volume fell 24.22% from 61,614 on 2025-07-01. youtube 3M was -14.29% from 7.0. For player ranks instead of growth, see the Steam Most Played API.

Longer windows still disagree by source

3M and 6M rows from the same calls keep the date split, which means chatgpt can rise 123.81% on youtube over 6M (47.0 from 21.0 on 2026-02-21) while google search only rises 8.22% over its own 6M (79.0 from 73.0) and wikipedia uses a third pair of dates entirely. chatgpt youtube 3M was +27.03% (47.0 from 37.0 on 2026-05-23). wikipedia chatgpt 3M used 2026-04-01 at 34.4 / 4,760 (+38.08%) and 6M used 2026-01-01 at 34.7 / 4,805 (+36.89%). google search 3M was -4.82% (79.0 from 83.0).

crocs amazon 3M was -7.07% (52.6 from 56.6, volumes 9,187,782 from 9,902,052) and 6M was +73.03% from 30.4 / 5,310,154. crocs youtube 3M was +58.93% (89.0 from 56.0) and 6M was +71.15% (89.0 from 52.0). Shopping 3M was -3.7% (52.0 from 54.0). stanley cup amazon 3M was +1.62% (18.8 from 18.5, volumes 2,263,655 from 2,230,295) and 6M was -23.89% from 24.7 / 2,974,958, while youtube 3M was -29.41% (12.0 from 17.0) and 6M was -7.69% (12.0 from 13.0). nike google search 3M was -15.79% (64.0 from 76.0) and 6M was +6.67% (64.0 from 60.0). For Amazon-only volume without the multi-source wrapper, see Amazon search trends.

Limits of a comma-separated growth call

Failed sources stay in the payload instead of aborting the whole call, which is why stanley cup on amazon, tiktok, youtube still returned amazon index 18.8 / 2,263,655 recent volume and youtube index 12.0 after tiktok came back unavailable here. tiktok returned unavailable while amazon and youtube succeeded, so sources_processed was 3 and sources_successful was 2. Retry tiktok as its own call, or drop it. news sentiment and news volume 504'd in this run and are the wrong sources to bury inside a list if the pipeline is down.

Volume-weighted growth is not a cross-channel average. crocs weighted -28.03 from amazon alone. chatgpt weighted +41.93 from wikipedia 6,574 page views, ignoring search index 79.0 and youtube index 47.0. Elden Ring weighted -19.46 with google search contributing 96.25% of total_volume via an estimated 1,200,000, next to steam's real 46,694 peak players at 3.75%. When volume_estimated is true, quote the 0-100 index and skip the derived volume.

Source-specific keyword formats do not relax inside a list. steam resolves the first Steam store search hit. wikipedia misses many parenthetical titles. amazon 7D/14D windows often collapse; this file used 30D and longer. wikipedia and steam recent_dates in this pull lagged into July (2026-07-01) while search and youtube sat on 2026-08-22 or 2026-08-15. For Wikipedia page-view growth without the wrapper, see the Wikipedia pageviews growth API. For YouTube search volume, see YouTube trends.

Pricing for multi-source growth

Free covers 100 requests per month across every source as of August 2026, and the product docs count get_growth per source plus keyword, so a three-source comma-separated string is three billed units even though the HTTP round trip is one. Starter is $19 for 1,000 requests, Pro $49 for 5,000, Business $199 for 25,000. Annual billing saves 20%. Check current pricing before locking a budget. Extra percent_growth periods inside that call do not add units. Confirm the billed count on the account page after a multi-source test rather than guessing from the JSON sources_processed field.