Source app rankings vs the Google Play board

An app rankings vs Google Play comparison on Trends MCP is two different tools. Source app rankings on get_growth or get_time_series returns a 0-100 series for one Android bundle ID, with chart-position volume where a lower number is a better slot. Type Google Play on get_top_trends returns ranked store display names and takes no keyword. On September 24, 2026 the Google Play board still carried as_of 2026-09-13T12:01:38 UTC, with WhatsApp Messenger at rank 1 and ChatGPT at rank 11. The matching bundle com.openai.chatgpt closed rank volume 1.0 and index 100.0 on 2026-09-23 across 30 daily points from 2026-08-25. Keyword ChatGPT returned HTTP 404 no_data. Mixing the two strings is a 400.

ChatGPT closed rank volume 1.0 next to Play rank 11

The September 24, 2026 pull is the split that still matters next month: live Play rank is a title list that can sit on an old stamp, while source app rankings keeps printing a different number for the same package. Omitting limit returned limit 25, count 25, offset 0, as_of_ts 2026-09-13T12:01:38.463213+00:00. Rank 1 was WhatsApp Messenger. Rank 3 was YouTube. Rank 5 was Free Fire: 9th Anniversary. Rank 11 was ChatGPT. Rank 21 was Google Gemini.

Live rank is not a sort of rankings volume. ChatGPT sat at Play rank 11 with rankings volume 1.0. Free Fire sat at Play rank 5 with the same volume 1.0 and the same index 100.0. Two titles can print "first" on the series while the board lists them at 5 and 11. Roblox printed index 100.0 at volume 2.0, so index 100.0 is not unique either. Quote both fields. Do not treat volume 1.0 as Play rank 1.

WhatsApp is the inverse. Play rank 1 closed rankings volume 164.0 and index 96.7 on 2026-09-23. get_time_series returned 7 daily rows from 2026-09-17 (index 0.0, volume 193) through 2026-09-23 (96.7, 164). 7D then compared 96.7 against that 0.0 baseline and printed growth 999999.0, while volume moved from 193.0 to 164.0, volume_growth -15.03%. Quote 164.0 next to Play rank 1. Skip 999999.0 in a forecast. 14D, 30D, 3M, and 12M all returned date_out_of_range with data_start 2026-09-17.

YouTube sat at Play rank 3 with rankings volume 199.0 and index 0.0, 15 points, recent_date 2026-09-23. 7D index growth was -100.0% from 47.9 on 2026-09-16, while volume worsened from 165.0 to 199.0, +20.61%. A Play-board top-three title can sit at the back of the rankings series. Gemini ran the other way: Play rank 21 next to rankings volume 6.0 and index 92.0. 14D volume improved from 10.0 to 6.0, volume_growth -40.0%, index +21.05% from 76.0. Rankings volume down is a better slot.

ChatGPT get_time_series printed 30 daily points, one row per date from 2026-08-25 through 2026-09-23. Volume stayed at 1 and index at 100.0 on 29 of those days. The exception is 2026-09-05, index 0.0 and volume 3. get_growth 7D, 14D, and 30D were then flat at 100.0 / 1.0. 3M and 12M returned date_out_of_range with data_start 2026-08-25, data_end 2026-09-23, 30 points. That window is about one month of daily grain, not a five-year weekly strip.

sort=rank_change with window=7d and limit 10 used the same stale as_of_ts, plus baseline_ts 2026-09-06T02:01:39.865148+00:00, then returned count 1. The only climber was YouTube Music at rank 150, prev_rank 151, rank_change 1. The board barely moved versus that baseline. offset 25 with limit 5 returned count 5 on the same stamp: Spotify: Music and Podcasts at 26, 8 Ball Pool at 27, Free Fire MAX at 28, Google at 29, Lazada 8.8 Sale at 30. The chart continued past row 25 on this pull.

TikTok 7D is the rank-volume warning in the other direction. Index fell from 100.0 to 70.0, -30.0%, while volume moved from 2.0 to 5.0, +150.0%. Index down and volume up can be the same worse slot. Snapchat 7D index rose 100.3% from 33.3 to 66.7 while volume improved from 20.0 to 19.0, -5.0%. Facebook 7D index rose 66.6% from 50.0 to 83.3 while volume improved from 5.0 to 3.0, -40.0%. Duolingo at Play rank 16 returned HTTP 404 no_data on source app rankings. A board title is not a guarantee the series exists.

Call source app rankings without pasting a display name as a type

Two strings, two tools. REST uses mode get_growth or get_top_trends. MCP uses the matching tool names. Rank-history windows live on the Android app rankings API. The live board page is the free Google Play trends API. Window defaults live on get-growth. The install-estimate split is documented with app downloads vs Google Play.

{"mode": "get_top_trends", "type": "Google Play"}
{"mode": "get_growth", "source": "app rankings", "keyword": "com.openai.chatgpt", "percent_growth": ["7D", "14D", "30D"]}
{"mode": "get_time_series", "source": "app rankings", "keyword": "com.openai.chatgpt"}
{"mode": "get_growth", "source": "app rankings, app downloads", "keyword": "com.openai.chatgpt", "percent_growth": ["7D"]}

The MCP equivalents are get_top_trends(type="Google Play"), get_growth(keyword="com.openai.chatgpt", source="app rankings", percent_growth=["7D", "14D", "30D"]), and get_time_series(keyword="com.openai.chatgpt", source="app rankings"). REST top_trends with type Google Play and limit 3 on September 24, 2026 returned the same snapshot timestamp as the omitted-limit MCP pull, with count 3: WhatsApp Messenger, Facebook, YouTube.

Omit percent_growth on com.openai.chatgpt after the three-window success returned HTTP 400, error growth_calculation_failed, message No growth window could be calculated. The default window is 12M. That preset printed date_out_of_range against data_start 2026-08-25. Store 7D, 14D, or 30D. Do not assume omit works because an earlier call on the same bundle did.

REST 1W matched 7D exactly on ChatGPT rankings: 100.0 from 100.0, volume 1.0 from 1.0, growth 0.0, baseline 2026-09-16, 30 points. 2W matched 14D: 100.0 from 100.0, baseline 2026-09-09. 1M matched 30D: 100.0 from 100.0, baseline 2026-08-25. Instagram first REST call with 2W and 14D returned HTTP 404 no_data. A retry then printed 93.2 from 91.5, +1.86%, volume 10.0 from 11.0, -9.09%, same recent_date 2026-09-23. Quote the 404. Retry. MCP had already returned that Instagram row on the first try.

The comma-separated ChatGPT call returned sources_requested ["app rankings", "app downloads"], sources_successful 2, and trend_analysis.uptrend_ratio 0/2. App rankings 7D was flat at 100.0 / 1.0 with 30 points and recent_date 2026-09-23. App downloads 7D was flat at 0.0 / 0.0 with 164 points and recent_date 2026-09-21. volume_weighted_growth printed 0.0, weight_coverage 1/2 sources, because the downloads close was 0.0 and the only volume used was rankings 1.0. Same preset, different grain, different close dates. Skip a trailing zero on downloads before joining it to rank history.

get_time_series on app rankings, app downloads returned invalid_source with 16 names and no all. Multi-source strings belong on get_growth only. Source App Rankings case-folded and returned HTTP 200, echoing the ChatGPT 7D row of 100.0 / 1.0. Store the lowercase source string in configs.

Mixing app rankings and Google Play is a 400

get_top_trends with type app rankings returned HTTP 400, error invalid_request, message Unknown type 'app rankings', plus 24 valid feed labels that include Google Play. get_growth with source Google Play and keyword com.openai.chatgpt returned HTTP 400, error invalid_source, message 'google play' is not a valid source., plus 16 valid sources: google search, google images, google news, google shopping, youtube, wikipedia, tiktok, reddit, amazon, news sentiment, news volume, app downloads, app rankings, steam, npm, python, all. The handler lowercased the rejected source. It did not map it to app rankings.

The inverse fold exists on the feed name. type google play on get_top_trends returned HTTP 200 and echoed type Google Play, with the same first three titles as the canonical label and as_of_ts 2026-09-13T12:01:38.463213+00:00. llms.txt still describes feed labels as case-sensitive. The live handler accepted this lowercase Play string on September 24, 2026. Prefer the canonical Google Play spelling in stored configs.

REST omitting type is a different 400. POST https://api.trendsmcp.ai/api with mode get_top_trends and limit 5, no type, returned HTTP 200 with envelope statusCode 400, error invalid_request, message The 'type' parameter is required, then the same 24-name list. llms.txt still names that miss missing_parameter. Read statusCode inside the JSON, not only the HTTP line.

country GB with limit 5 on Google Play returned HTTP 400, error invalid_request, message that country is only supported for Google Trends, Google Trends by Category, TikTok Trending Hashtags, TikTok Trending Hashtags by Category, X (Twitter) Trending, and YouTube Trending. Google Play is not on that list. Omit country on this feed. Rank 7 Flipkart Online Shopping App and rank 30 Lazada 8.8 Sale still appeared on the default board. This chart is not a single-country consumer list.

GitHub and Google Trends REST pulls on the same morning stamped as_of_ts 2026-09-24T06:01:41.734023+00:00. Google Play did not. The Play board on this run is eleven days behind those other feeds, while app rankings for ChatGPT closed 2026-09-23. Quote the board stamp next to the series recent_date. Do not join them as the same morning.

get_top_trends MCP pricing is 1 credit per 10 requested rows, max 10 credits per call. REST counts 1 request per call. Free boards still cap at the top 10. The 25-row pulls on this page are the paid default, which is also what the handler returns when limit is omitted on a paid key. Rank-history follow-up stays on source app rankings. This split stays inside one store: a bundle-ID rank series versus live chart rank.