Why does rank_change omit Muse?

On 2026-09-24 the rank sort listed Muse from Meta at 1. sort=rank_change window 7d used baseline 2026-09-17 and returned climbers such as Capital One Shopping at rank 17 from 73. A title that is already number one has no rank_change row. Read the default rank sort to confirm it is still on top.

What is the Android bundle ID?

com.facebook.aura. That is the id= value on the Play listing titled Muse from Meta by Meta Platforms, Inc., 4.7 stars, 16.5K reviews, updated 2026-09-21. Display names fail on source app rankings. Keyword Muse from Meta is the App Store chart label, not the rankings keyword.

Does omitting percent_growth work for muse ai?

On google search, yes. Omit defaults to 12M, and this pull printed 26 from 4, +550.0%, the same row as 1Y, 261 points. On app rankings for com.facebook.aura, 30D already returned date_out_of_range because data_start is 2026-09-12. Send 7D and 14D on that source. Do not assume the search default applies to a two-week Android series.

Connecting Cursor is free. Calls are not. Free is 100 requests per month. Starter is $19 for 1,000. get_growth bills one request per source plus keyword. get_top_trends on MCP costs 1 credit per 10 requested rows, max 10 credits. REST bills 1 request per get_top_trends call. A proving trio is three requests.

How to track Muse from Meta in Cursor

Three Trends MCP calls cover Muse from Meta: the iOS free chart, the muse ai search series, and Android rank history on com.facebook.aura. On September 24, 2026, get_top_trends type App Store Top Free listed Muse from Meta at rank 1, stamp 2026-09-24T12:01:38 UTC. Dedicated get_growth on source google search, keyword muse ai, closed 26 on 2026-09-19, +333.33% over 30D and +550.0% over 1Y, with a 7D drop of 39.53% from 43. Source app rankings on com.facebook.aura closed index 100.0 and volume 1.0 on 2026-09-23, 12 daily points from 2026-09-12. sort=rank_change skips that number one. Type muse ai is a 400.

The three calls that actually track Muse

Store three strings and stop mixing them. The live iOS label is Muse from Meta. The search and YouTube keyword is muse ai. The Android rankings keyword is com.facebook.aura. Meta's 2026-09-08 post, Introducing Muse, is the product description. It is not an API argument.

{"mode": "get_top_trends", "type": "App Store Top Free", "limit": 20}
{"mode": "get_growth", "source": "google search, youtube", "keyword": "muse ai", "percent_growth": ["7D", "30D", "1Y"]}
{"mode": "get_growth", "source": "app rankings", "keyword": "com.facebook.aura", "percent_growth": ["7D", "14D"]}

MCP equivalents: get_top_trends(type="App Store Top Free", limit=20), get_growth(keyword="muse ai", source="google search, youtube", percent_growth=["7D", "30D", "1Y"]), get_growth(keyword="com.facebook.aura", source="app rankings", percent_growth=["7D", "14D"]). REST is POST https://api.trendsmcp.ai/api with Authorization: Bearer YOUR_API_KEY. Cursor Agent needs the phrase "using TrendsMCP" or it will web-search.

The September 24 chart pull returned count 20: Muse from Meta 1, ChatGPT 2, Google Gemini 7, Claude by Anthropic 12, WhatsApp Messenger 15. Search close for muse ai was 26 from 43 on 7D, 6 on 30D, 4 on 1Y, 261 weekly points, recent_date 2026-09-19. YouTube on the comma list closed 31, +34.78% 7D from 23, +342.86% 30D from 7, +520.0% 1Y from 5. That YouTube-versus-search split is documented on Muse Google Search vs YouTube. The four-app search table is on Muse vs ChatGPT.

Omit percent_growth on muse ai / google search and the default 12M reprints 26 from 4, +550.0%, baseline 2025-09-20. Same numbers as 1Y on this pull. Keep sending 7D anyway. The 7D drop is the part a 12M default hides.

Rank change will skip Muse at number one

sort=rank_change, window=7d, type App Store Top Free, used the same as_of_ts as the rank board plus baseline_ts 2026-09-17T02:01:38.593365 UTC. count 25. Muse from Meta is not in the list. ChatGPT is not in the list. The leader of that mover board was Capital One Shopping: Save Now, rank 17, prev_rank 73, rank_change 56. Grok AI appeared at 31 from 37, rank_change 6. A monitor that only stores rank_change will report Muse as missing on the same morning the rank sort still prints it at 1.

Default sort is rank. Use that to answer "is Muse still on top." Use rank_change to answer "what entered or climbed." Quote baseline_ts next to as_of_ts. The App Store Top Free API page is the feed contract. This page is the Muse-shaped workflow.

Google Play is the wrong live board for this launch on this pull. Type Google Play, limit 25, still stamped 2026-09-13T12:01:38.463213 UTC when the App Store board stamped 2026-09-24. Play rank 1 was WhatsApp Messenger. ChatGPT was 11. Gemini was 21. No Muse row in 25. The Play listing for com.facebook.aura exists: Muse from Meta, Meta Platforms, Inc., 4.7 stars, 16.5K reviews, last updated 2026-09-21. A listing page is not a get_top_trends row. Quote the frozen as_of_ts if Play is cited at all.

get_top_trends with type muse ai returned HTTP 400, error invalid_request, message Unknown type 'muse ai', and listed 24 valid types including App Store Top Free and Google Play. get_growth with source Google Play and keyword muse ai returned HTTP 400, error invalid_source, message 'google play' is not a valid source., plus 16 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 rejected source was lowercased. It was not mapped.

Android history on com.facebook.aura is short

Source app rankings, keyword com.facebook.aura, windows 7D, 14D, 30D, returned 12 points, data_start 2026-09-12, data_end 2026-09-23. Recent index 100.0, recent volume 1.0. 7D and 14D compared that 100.0 with baseline index 0.0 and printed growth 999999.0, while volume moved from 2.0 to 1.0, volume_growth -50.0%. 30D returned date_out_of_range with the message that preset 30D needs a baseline outside available data. Skip 999999.0. Quote volume 1.0 next to index 100.0. Send 7D and 14D only until the series is older than a month.

Source app downloads on the same bundle is thinner and should not be treated as a demand crash. 7 points, data_start 2026-09-10, data_end 2026-09-22. Recent value 0.0, recent volume 0.0. 7D compared that zero with 36.3 / 39907 on 2026-09-15 and printed -100.0%. 14D compared with 0.4 / 433 on 2026-09-10, also -100.0%. 30D date_out_of_range. A seven-row downloads strip that closes at 0.0 is incomplete coverage. Prefer the rankings volume 1.0 and the App Store rank until downloads has more than a handful of dates.

ChatGPT's bundle com.openai.chatgpt on app rankings is the contrast: 30 points from 2026-08-25 through 2026-09-23, index 100.0 and volume 1.0, 7D/14D/30D growth 0.0. Gemini's com.google.android.apps.bard closed 92.0 / 6.0 on 2026-09-23, 30 points, 7D flat. Two packages can print index 100.0. Volume is the slot hint: Muse 1.0, Gemini 6.0. Join iOS rank on the display name, Android rank on the bundle.

Bare muse on google search closed 47 with estimated volume 815,000 on 2026-09-19. That is the band. muse from meta on google search closed 23, data_quality low, over 90% zeros, 30D and 1Y at 999999.0. Do not save either string as the Muse tracker keyword.

Cursor mcp.json for this workflow

Cursor stores Trends MCP as url plus transport http under mcpServers, pointed at the api host. Mac and Linux: ~/.cursor/mcp.json. Windows: %USERPROFILE%\.cursor\mcp.json. Fully quit Cursor after saving.

"trends-mcp": {
  "url": "https://api.trendsmcp.ai/mcp",
  "transport": "http",
  "headers": {
    "Authorization": "Bearer YOUR_API_KEY"
  }
}

Do not paste Claude Desktop's npx mcp-remote block into this file. Do not paste the www OAuth MCP host here; that endpoint is for Claude.ai and will 401 in Cursor. Windsurf wants serverUrl. VS Code wants type http under servers. The host reference is Trends MCP for Cursor.

Free 100 requests per month. Starter $19 / 1,000. Pro $49 / 5,000. Business $199 / 25,000. Annual 20 percent off. A proving trio is one App Store page, one muse ai growth row, one com.facebook.aura rankings row. Three requests. Adding youtube on the comma growth call is a second source on that keyword, so it adds a credit. Paging limit 50 four times on the chart spends four requests to reprint rank 1.

Cron does not belong in Cursor. Nightly boards go to REST at https://api.trendsmcp.ai/api with the same Bearer key and mode get_top_trends or get_time_series. Agent only runs when the IDE is open. Connecting is free. Calling is not.

If tools miss after a full quit, check the host before blaming Muse. www in mcp.json is the usual miss. If the chart row says Muse from Meta and growth on that exact string looks empty, switch the keyword to muse ai. If rankings 404, check the bundle on the Play URL, not the iOS caption.