Google Trends

What Google Trends MCP server does

Google Trends MCP server exposes Google Trends research through MCP tools. It can retrieve current Trending Now data for countries and regions, including search volume, growth, start time, category, status, and related queries. It can also inspect the news and hourly activity associated with an individual trend.

Keyword research includes interest over time, interest by country or subregion, related queries, related topics, and keyword overviews. Queries can use different locations, categories, search types, and time ranges. Supported search types include Web, YouTube, News, Images, and Shopping.

The server also provides analysis tools for momentum, seasonality, content scheduling, share of search, multi-term rankings, spikes, daily history, and keyword ideas. Local history features can save Trending Now results, maintain watchlists, and report changes over time.

How it works

The server runs locally and communicates with an MCP client over stdio or HTTP. Requests are sent to Google Trends, then returned as structured results for the connected agent to interpret. No API key, Google Cloud project, browser automation, or sign-in is required.

Google Trends MCP server uses paced requests, a persistent session, back-off behavior, and caching. The cache can provide results when Google does not respond, while the response indicates that cached data was used. Trend and chart requests use separate budgets so a limit in one area does not stop the other.

Google Trends values are relative 0–100 interest indexes, not absolute search counts. Partial time periods are identified, and failed requests return an error reason with guidance rather than an empty result.

Setup and configuration

The project supports Python 3.10 through 3.13 and is tested on Linux, macOS, and Windows. It is available as the gtrends-mcp-full Python package and is configured as a local MCP server in a compatible client. The provided material does not specify a required environment variable or client-specific configuration value.

No API key or sign-in configuration is needed. The server uses two dependencies and stores local history, snapshots, and watchlist data on the machine. Data requests go to Google; local persistence features use local files.

Tools and capabilities

The available tools cover five practical areas:

• Current trends: Find live trends, inspect their details, retrieve country feeds, compare activity across countries, and match trends to subjects or brands.
• Keyword exploration: Review interest over time, geographic interest, related searches, related topics, period comparisons, and location comparisons.
• Analysis: Classify momentum, identify seasonal peaks, create content calendars, calculate share of search, rank many terms, detect spikes, inspect daily history, and generate keyword ideas.
• Local history: Save Trending Now snapshots, search saved trends, add or remove watchlist terms, and produce watchlist reports.
• Lookup and diagnostics: Find Google Trends location codes, identify categories and topics, and check endpoint failures.

The tools accept common natural-language research requests through the MCP client. For example, an agent can compare terms over several years, find regions where one term leads another, or identify when seasonal content should be published.

Limitations and notes

Google Trends data is normalized rather than a direct measure of search volume. A 100 index value represents the highest relative interest in the selected comparison, location, and period; it does not mean 100 searches. Results may also be partial for incomplete periods.

The website's normal limits do not apply to every server feature: the project states that comparisons can include up to 25 terms, while standard interest-over-time comparisons support up to five terms. Daily history and saved Trending Now data extend beyond the shorter history available on the website, but saved history only exists from the time the server begins recording it.

The project is MIT licensed. It is intended for MCP clients such as Claude Desktop, Claude Code, Cursor, Windsurf, VS Code with Copilot agent mode, and OpenAI Codex CLI.