Integrations
Claude calls get_growth after Trends MCP is added on claude.ai, Desktop, or Claude Code. The proving call is a keyword percent window, not a trending board.
Claude pulls keyword percent change through Trends MCP get_growth after the connector is added. claude.ai uses OAuth at https://www.trendsmcp.ai/mcp with no API key in the connector. Desktop uses npx mcp-remote against https://api.trendsmcp.ai/mcp with a Bearer key. The homepage fetched 2026-11-16 still tells Claude Code to add the www host and sign in with /mcp. llms.txt and Trends MCP for Claude Code from the same morning still add the api host with a Bearer header. Ask, using TrendsMCP, for a keyword source such as Google Search plus a phrase. Omit percent_growth and the window is still 12M. A 2026-11-16 nike Google Search call printed index 58.0 on 2026-11-14 versus 65.0 on 2025-11-15, growth -10.77 percent, across 261 weekly points.
Host setup for the browser and Desktop lives on Trends MCP for Claude. The terminal add lives on Trends MCP for Claude Code. Those pages prove the connect with get_top_trends. This page is the growth-shaped call those guides never name.
claude.ai: Customize, Connectors, then Add custom connector.
https://www.trendsmcp.ai/mcpThat www URL is intentional. Claude handles OAuth. No API key belongs in the connector. Deep link: https://claude.ai/customize/connectors?modal=add-custom-connector&connectorName=Trends%20MCP&connectorUrl=https%3A%2F%2Fwww.trendsmcp.ai%2Fmcp. Start a new chat after consent. An old thread will not see a connector added mid-conversation.
Anthropic's custom-connector help, fetched 2026-11-16 from https://support.claude.com/en/articles/11175166-get-started-with-custom-connectors-using-remote-mcp, now lists Claude, Cowork, and Claude Desktop on Free, Pro, Max, Ultra, Team, and Enterprise. Free users are limited to one custom connector. Pro and Max add from Customize, Connectors, "+ Add", then "Add custom connector". Team and Enterprise owners add the server under Organization settings first; members later click Connect. The dedicated Claude page still says Pro, Max, and Team are the usual remote-connector surfaces. Recheck Anthropic's matrix before a rollout that assumed Free was blocked. This run did not click Add custom connector and did not complete OAuth in a Claude tab. It recorded the published steps.
Anthropic also notes that a remote connector talks to the MCP host from Anthropic's cloud on HTTPS port 443, not from the laptop. Local claude_desktop_config.json servers are a separate path and do use the local network. Trends MCP is a public HTTPS host, so the remote recipe is the one claude.ai needs.
Desktop talks to the api host. User, Settings, Developer, Edit Config, inside mcpServers:
"trends-mcp": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://api.trendsmcp.ai/mcp", "--header", "Authorization:${AUTH_HEADER}"],
"env": { "AUTH_HEADER": "Bearer YOUR_API_KEY" }
}
Mac path: ~/Library/Application Support/Claude/claude_desktop_config.json. Windows path: %APPDATA%\Claude\claude_desktop_config.json. Key from the account page, format tmcp_live_. Fully quit Desktop after saving. A window reload is not enough.
llms-full.txt fetched 2026-11-16 inlines the Bearer string in args and drops the env map. The dedicated Claude page keeps Authorization:${AUTH_HEADER} plus AUTH_HEADER. Both hit the same api host. Do not paste Cursor's url plus headers object into this file. Pasting it into Desktop is the wrong shape.
Ask with "using TrendsMCP" so Claude routes to the server instead of a web search. The proving prompt is a growth call, not a board call.
get_growth(keyword="nike", source="google search")
get_growth(keyword="nike", source="google search, youtube", percent_growth=["3M", "12M"])
get_top_trends with type Google Trends still ranks a live board. That is a different tool. Do not send a board type string into get_growth. The dedicated Claude page still tells the model to call get_trends for keyword history. The live MCP list in this session advertised get_time_series, get_growth, get_top_trends, get_context, get_reviews, and get_usage. Six tools. Five of them bill. get_usage is free. get_top_posts is absent. get_ranked_trends is absent.
Nightly growth tracking belongs on REST, not a Claude loop that only runs when the chat is open. POST https://api.trendsmcp.ai/api with a Bearer key and mode get_growth. Wrapping claude in GitHub Actions just to pull windows is a roundabout way to hit that URL.
The homepage fetched 2026-11-16 tells Claude Code operators to run this add, then type /mcp to sign in:
claude mcp add --transport http --scope user trends-mcp https://www.trendsmcp.ai/mcp
The same homepage groups Claude Code with Claude and ChatGPT as sign-in clients on https://www.trendsmcp.ai/mcp. Cursor and VS Code stay on the api host with a Bearer key.
llms.txt last updated 2026-11-16 still publishes the other recipe:
claude mcp add --transport http trends-mcp https://api.trendsmcp.ai/mcp --header "Authorization: Bearer YOUR_API_KEY"
docs.md and llms-full.txt from the same morning repeat that api-host line. The dedicated Claude Code page repeats it too, and says not to paste https://www.trendsmcp.ai/mcp into the CLI, because www is OAuth for claude.ai and a www add 401s. This run did not execute claude mcp add. It recorded both published commands. Match the recipe to the binary that is actually open. A Desktop npx block does not install the CLI. A claude.ai connector does not install the CLI.
CLAUDE.md can name get_growth, keyword sources such as Google Search, YouTube, and Amazon, and a cap of one source then stop. It should not freeze this morning's nike indexes. Those points move. A frozen percent in CLAUDE.md becomes a stale architecture note.
Enable the connector per chat from the "+" control, then Connectors, if the thread was opened before the add. Anthropic's help describes that toggle. A green connector row in settings without a tool result is only a settings entry.
A 2026-11-16 MCP get_growth call that sent source Google Search and keyword nike with no percent_growth field printed one 12M row and metadata.calculations_completed 1. Recent index 58.0 on 2026-11-14 00:00:00+00:00. Baseline 65.0 on 2025-11-15 00:00:00+00:00. Growth -10.77 percent, direction decrease, 261 weekly points. volume_available true, volume_estimated true, recent_volume 26,500,000, baseline_volume 29,600,000. volume_growth was null, with volume_growth_omitted_reason set to "volume is derived from the trend value for this source, not an independent measurement". Quote the 0-100 value. Treat that volume label as derived.
A second call that sent six presets completed 6 calculations on the same 261 points. Alias pairs that shared dates and growth on that nike series:
| Labels | Baseline date | Baseline index | Growth |
|---|---|---|---|
| 7D and 1W | 2026-11-07 | 53.0 | +9.43% |
| 12M and 1Y | 2025-11-15 | 65.0 | -10.77% |
Recent index was 58.0 on 2026-11-14 for every row. Windows that did not collapse into those aliases still shared that recent point. YTD was 0.0 percent from 58.0 on 2026-01-03, direction flat. 3M was -13.43 percent from 67.0 on 2026-08-15. Same 58.0, four different baselines. For comma-separated sources in one get_growth body, see the multi-source growth API. The shared window table lives on get_growth.
Preset strings on MCP are 7D, 1W, 14D, 2W, 30D, 1M, 2M, 3M, 6M, 9M, 12M, 1Y, 18M, 24M, 2Y, 36M, 3Y, 48M, 4Y, 60M, 5Y, MTD, QTD, and YTD. Named {name, recent, baseline} objects are REST-only. REST mode growth with source Google Search, keyword nike, and percent_growth ["12M"] printed the same 58.0 versus 65.0 row. A REST body that sent {"name": "yoy", "recent": "2026-10-31", "baseline": "2025-11-01"} printed 52.0 versus 53.0, growth -1.89 percent, plus recent_requested and baseline_requested. Those snap fields are the REST custom-date contract. Do not send a date object to MCP and expect them.
MCP source Google Trends on keyword nike returned 400 invalid_source, message 'google trends' is not a valid source. The valid_sources array named google search, google images, google news, google shopping, youtube, wikipedia, tiktok, amazon, app downloads, app rankings, steam, npm, python, pinterest, and all. That list is the keyword-history contract. Type Google Trends belongs on get_top_trends. Keep the board call on that type string.
A 2026-11-16 get_growth call that sent source Reddit and keyword nike returned 400 source_disabled, message 'reddit' is temporarily disabled for get_growth. Try another source. The same valid_sources list followed. news sentiment used the same source_disabled code. OpenAPI's Error.error enum, fetched the same morning, still lists invalid_source and growth_calculation_failed. It does not list source_disabled. A generated client that only maps that enum will not name the failure Claude just hit.
Amazon is still a growth source. Keyword air fryer on source Amazon with windows 7D, 1M, and 12M printed 52 monthly points and metadata.all_successful false. 7D returned error collapsed_range with message "Recent and baseline resolved to the same data point; not enough history for this preset". 1M succeeded on 2026-10-31 versus 2026-09-30: index 67.9 versus 59.7, +13.74 percent, volumes 7,350,631 versus 6,465,752, volume growth +13.69 percent. 12M was -5.83 percent from 72.1 / 7,802,064 on 2025-10-31. A Claude prompt that mixes Amazon 7D with Google Search 7D will look broken on Amazon and fine on Search. Read status per row.
The live google-search-vs-reddit page still documents source reddit on get_growth as a working daily community series. That 400 is the current contract. Do not send Claude a Project instruction that names reddit as a keyword source until the disable lifts.
Docs price get_growth per source in the call, every window included. The six-window nike Google Search call and the default 12M call are both one unit. Extra percent_growth strings do not add units. source all is the expensive exception: the MCP schema says that alias expands to 9 keyword sources and costs 9. Do not ask Claude to "compare every platform" in one sitting.
A comma list is the other meter. source Google Search, YouTube with windows 12M and 3M printed requests_charged 2, sources_successful 2, and growth_specs_count 2. Google Search matched the 12M -10.77 percent and 3M -13.43 percent rows above. YouTube printed 261 points, 12M 42.0 versus 44.0 on 2025-11-15 (-4.55 percent), and 3M 42.0 versus 43.0 on 2026-08-15 (-2.33 percent). YouTube rows in this pull omitted volume fields. volume_weighted_growth used Google Search alone: -10.77, sources_with_volume 1, weight_coverage 1/2 sources. Read source_results before treating that mixer as a two-platform number.
Connecting is free. Calling is not. get_usage during this run printed Business 7,381 of 25,000 used at the start, then 7,424, reset 2026-12-01. cost_rule: 1 request per data type in a call; get_top_trends: 1 request per 10 rows per board (limit defaults to 100). Only data types that answer are charged. buy_credits_url on that payload still uses the apex account billing tab:
https://trendsmcp.ai/account?tab=billing
Site links on this page stay on www. pricing.md fetched 2026-11-16 already publishes the www billing tab.
A garbage REST mode getgrowth returned envelope statusCode 400, error invalid_mode, and valid_modes get_context, get_growth, get_reviews, get_time_series, get_top_trends. The aliases object named growth for get_growth and reviews for get_reviews. It omitted get_context. REST still serves the dropped posts mode. POST https://api.trendsmcp.ai/api with mode get_top_posts, source tiktok, keyword fitness, and limit 1 returned HTTP 200, count 1, and https://www.tiktok.com/@treandsara/video/7705638158607682829, posted 2026-11-08, 2,300,000 plays, 134,500 likes. Claude that reads the live MCP list will not offer that call. The existing get-top-posts-in-claude page still tells Claude to call it.
OpenAPI 3.1 at https://www.trendsmcp.ai/openapi.json, fetched 2026-11-16, describes GetGrowthRequest with alias growth and REST-only custom date objects. components.schemas still has no GetReviewsRequest. The mode discriminator mapping lists get_context, get_growth, get_time_series, and get_top_trends only. GetTopTrendsRequest still defaults limit to 25. MCP and get_usage cost_rule default the board to 100. HTML /docs grepped 1 leftover get_top_posts comment and 0 for GetReviewsRequest.
Free plan 100 requests a month. Starter $19 for 1,000. Pro $49 for 5,000. Business $199 for 25,000. One Google Search call plus one YouTube call is two requests. Looping source all finishes the free plan before the Project instruction is written. Check current pricing before locking a budget.
FAQ