Wikipedia Trends MCP

What Wikipedia Trends MCP MCP server does

Wikipedia Trends MCP MCP server gives MCP-compatible agents access to trend queries through the TrendsMCP service. Its primary use case is examining Wikipedia page-view activity over time, while the underlying service also documents multiple trend sources and live feeds. The server exposes three tools rather than requiring an agent to scrape pages or manage separate platform credentials.

get_time_series returns historical observations for one keyword on one source. Results can be weekly by default, or daily where the REST interface supports that mode. The returned index uses a 0–100 scale when the source pipeline supports normalization, and an absolute volume series may also be available.

get_growth calculates point-to-point percentage change. A request can include several comparison windows, such as 7D, 6M, 12M, 1Y, or 5Y, and those windows count as one request for the selected source and keyword.

get_top_trends returns a live ranked list and does not accept a keyword. It supports pagination, result limits, rank or rank-change sorting, and category or time-window options for applicable feeds.

How it works

The hosted MCP endpoint is https://api.trendsmcp.ai/mcp. Requests authenticate with a TrendsMCP Bearer key. The same key is used with the documented REST endpoint at https://api.trendsmcp.ai/api, although REST parameter behavior differs in a few places.

For MCP calls, get_time_series and get_growth require both a source and keyword, and each call targets one source except that growth accepts a comma-separated source list. get_top_trends requires a type value on MCP. That value must exactly match the feed name, including capitalization. REST can omit type to request every feed, but billing applies per feed in that case.

The repository also provides a stdio adapter. It can list tools without a key, but calls that access paid data require the configured API key and use the same quota as hosted requests. Wikipedia Trends MCP MCP server therefore fits clients that support remote HTTP MCP, streamable HTTP, or a local process adapter.

Setup and configuration

Create a TrendsMCP account and obtain an API key before making data requests. For hosted clients, configure the MCP URL and send the key as an Authorization: Bearer YOUR_API_KEY header. The README provides client-specific examples for Cursor, VS Code, Windsurf, Cline, Claude Code, and Claude Desktop through mcp-remote.

For local use, install the repository and start its module:

pip install -e .
python -m trends_mcp_server

Set TRENDSMCP_API_KEY in the stdio client's environment. Claude.ai's custom connector uses the hosted URL with OAuth instead of placing a Bearer key in the connector configuration.

Tools and capabilities

• get_time_series: Retrieves weekly or daily history for one source and keyword.
• get_growth: Reports percentage change using preset or custom date windows.
• get_top_trends: Lists current ranked items, with optional limits, offsets, categories, sorting, and supported comparison windows.
• REST compatibility: Uses mode: "get_time_series"; REST also accepts get_trends as an alias.
• Normalized values: Returns a 0–100 index where the selected pipeline supports it.

Limitations and notes

Each time-series request covers one source and keyword. Feed names for get_top_trends must match exactly on MCP, and some feed types require a category. Its default limit is 25 and the maximum is 200. Growth presets include multiple day, month, quarter, year, and year-to-date windows, but the available data shape depends on the selected source.

The README advertises a free allowance of 100 requests per month. The repository is licensed under MIT. No client-specific support is documented for ChatGPT in the allowed client list, although the README includes a ChatGPT connector setup path.