Call google search and google shopping on the same phrase, then read estimated Search volume next to a Shopping index that has no volume key.
Live data as of 2026-09-05
A Google Search vs Google Shopping comparison on Trends MCP is two get_growth sources on the same phrase. Source google search returns a weekly 0-100 index and, on some keywords, an estimated volume derived from that index. Source google shopping returns the weekly shopping-intent index with no volume key. From the September 5, 2026 pull, air fryer Shopping 7D rose 33.33% from 6.0 to 8.0 while Search fell 4.92% from 61.0 to 58.0. Omitting percent_growth on Search still returned period 12M (+13.73%). get_time_series rejects a comma-separated source list. Google Shopping is not a get_top_trends type. The free plan includes 100 requests per month.
The September 5, 2026 7D window is the split that still matters next month: purchase-intent interest can rise in the same week that general Search interest falls. Air fryer Shopping moved from 6.0 on 2026-08-15 to 8.0 on 2026-08-22. Search moved from 61.0 to 58.0 on those same dates. Passing 1W next to 7D returned the same recent_date, baseline_date, values, and percents on both sources. The MCP schema lists 1W as an alias.
Longer windows reverse the short-week story. Air fryer Shopping 12M was +60.0% (5.0 on 2025-08-23 to 8.0). Search 12M was +13.73% (51.0 to 58.0) with estimated recent_volume 4,450,000. Shopping 3M was -84.91% from 53.0 on 2026-05-23, the week after Shopping hit index 100.0 on 2026-05-16. Search 3M was -18.31% from 71.0, and Search had already peaked at 100.0 a week earlier on 2026-05-09 with series volume 7,680,751. The two series share a Saturday grain. They do not share a peak week.
chatgpt shows the other polarity. Search YTD was +22.95% (61.0 on 2026-01-03 to 75.0 on 2026-08-29). Shopping YTD was -34.48% (58.0 to 38.0). Brand curiosity can climb while shopping-intent interest for the same string cools.
Two source strings, one keyword. REST uses mode get_growth. MCP uses the get_growth tool. The shopping-only page is the Google Shopping trends API. The Search keyword page is Google search data. The get_growth contract page covers default windows.
{"mode": "get_growth", "source": "google search", "keyword": "air fryer", "percent_growth": ["7D", "1W", "30D", "3M", "12M", "YTD"]}
{"mode": "get_growth", "source": "google shopping", "keyword": "air fryer", "percent_growth": ["7D", "1W", "30D", "3M", "12M", "YTD"]}
{"mode": "get_growth", "source": "google search, google shopping", "keyword": "air fryer", "percent_growth": ["7D", "12M"]}
{"mode": "get_growth", "source": "google search", "keyword": "air fryer"}
The last call omitted percent_growth. The response still used period 12M, recent_value 58.0, baseline_value 51.0, and growth 13.73. That matches the 12M row inside the multi-window Search block.
The comma-separated air fryer call returned sources_requested ["google search", "google shopping"], sources_processed 2, sources_successful 2, plus a trend_analysis block with uptrend_ratio 1/2. That envelope is not two stacked single-source payloads. Read source_results for the per-source windows. metadata.total_data_points was 260 on every Search and Shopping block in this pull, including nike, running shoes, and chatgpt.
| Keyword | Search 12M | Search recent | Shopping 12M | Shopping recent | Close date |
|---|---|---|---|---|---|
| air fryer | +13.73% (51.0 to 58.0) | 4,450,000 est. | +60.0% (5.0 to 8.0) | no volume | 2026-08-22 |
| running shoes | +6.67% (30.0 to 32.0) | no volume field | +40.0% (5.0 to 7.0) | no volume | 2026-08-29 |
| nike | -2.94% (68.0 to 66.0) | 27,500,000 est. | -12.5% (32.0 to 28.0) | no volume | 2026-08-29 |
| chatgpt | -17.58% (91.0 to 75.0) | no volume field | -7.32% (41.0 to 38.0) | no volume | 2026-08-29 |
Air fryer closed a week earlier than nike, running shoes, and chatgpt on the same day. Quote recent_date per keyword. Do not assume one Saturday for the whole pull.
Amazon search is a third commerce calendar, monthly, with true volume. That pairing lives on Google Shopping vs Amazon search. Mixing Search and Shopping first keeps both series weekly.
google search volume is not a second measurement. On air fryer and nike, get_growth set volume_available true and volume_estimated true, then set volume_growth null with volume_growth_omitted_reason "volume is derived from the trend value for this source, not an independent measurement." Quote the index move. Treat the volume integer as a scaled reading of that index.
The time-series row is more precise than the growth round. Air fryer Search closed at value 58.0 with volume 4,454,836 on 2026-08-22. The matching get_growth window printed recent_volume 4,450,000. Nike Search closed at value 66 with volume 27,504,603 on 2026-08-29. get_growth printed 27,500,000.
running shoes and chatgpt Search windows in this pull had no volume_available field at all. A dedicated running shoes 12M call still omitted it. The comma-separated running shoes call then set has_volume_data false, volume_weighted_growth null, and weight_coverage 0/2 sources. Air fryer's weighted print was -4.92 with weight_coverage 1/2 sources and volume_percentage 100.0 on google search. That -4.92 is the Search 7D move. Shopping's +33.33% 7D did not enter the weighted figure because Shopping has no volume to weight.
google shopping get_time_series rows for air fryer used keys date, value, keyword, and source. There was no volume key to null out. Do not fill one in.
get_growth accepts google search, google shopping. get_time_series does not. The air fryer call with that comma-separated string returned HTTP-style status 400, error invalid_source, and message 'google search, google shopping' is not a valid source. The valid_sources list held 16 names: google search, google images, google news, google shopping, youtube, wikipedia, tiktok, reddit, amazon, news sentiment, news volume, app downloads, app rankings, steam, npm, python.
Two series calls are the chart path. Air fryer Search ran 260 weekly points from 2021-09-04 (value 33.0, volume 2,534,648) through 2026-08-22 (value 58.0, volume 4,454,836). Air fryer Shopping ran 260 points on the same Saturday grid, from 5.0 to 8.0, with no volume field. Nike Search started a week later (2021-09-11, value 67, volume 27,921,339) and closed a week later (2026-08-29, value 66, volume 27,504,603). Same tool, same source string, different close.
REST mode get_time_series aliases (get_trends, trends, time_series, lookup) still take one source. Prefer the canonical name in new code.
get_top_trends with type Google Shopping and limit 1 returned error invalid_request and message Unknown type 'Google Shopping'. The payload listed 24 valid feeds, including Google Trends and Google Trends by Category. It did not list a Shopping board.
Discovery of breakout queries uses type Google Trends. History of shopping-intent interest for a chosen phrase uses source google shopping. History of general Search interest uses source google search. Mixing the live type string into get_growth is the usual invalid_source. Mixing the keyword source string into get_top_trends is the usual invalid_request.
Paid plans start at Starter $19 per month for 1,000 requests. Check current pricing before a production budget. Failed 4xx calls do not count against the monthly quota.
FAQ