News Sentiment Trends MCP

What News Sentiment Trends MCP MCP server does

News Sentiment Trends MCP MCP server exposes trend data to MCP-compatible AI clients. It connects agents to Google, TikTok, YouTube, Amazon, Reddit, and more than 30 other listed data sources through one API key. The service supports keyword-based history and growth queries, plus live ranked feeds that do not require a keyword.

The main use cases are comparing interest in one topic across sources, checking how a topic changed between two points, and retrieving what is currently ranking on a supported platform. It is intended for agents that need structured trend data rather than browser scraping.

How it works

The hosted MCP endpoint is https://api.trendsmcp.ai/mcp. Requests use a Bearer key. The same key can be used with the REST endpoint at https://api.trendsmcp.ai/api, where the request mode identifies the operation.

get_time_series returns weekly history by default, with daily history available through REST. Each call accepts one source and one keyword. Where the source pipeline supports it, values use a 0–100 index, with 100 representing the peak in the returned period. An absolute volume series may also be available.

get_growth calculates point-to-point percentage change. It accepts one source or a comma-separated source list and can return several comparison windows in one call. Supported presets include 7D, 30D, 6M, 12M, 1Y, 2Y, 5Y, MTD, QTD, and YTD, along with custom date objects.

get_top_trends returns a live ranked list without a keyword. MCP requests require a type value that exactly matches the feed name, including capitalization. Results can be limited, paginated, sorted by rank or rank change, and filtered by category for some feed types.

Setup and configuration

Create an account at TrendsMCP and obtain an API key. For hosted clients, configure the MCP URL and send the key as Authorization: Bearer YOUR_API_KEY. Claude.ai custom connectors use the hosted URL with OAuth instead of placing a Bearer key in the connector configuration.

The repository also provides a Python stdio adapter. Install the package from the repository, then run python -m trends_mcp_server. Set TRENDSMCP_API_KEY in the process environment for paid calls; the adapter can list tools without a key, but requests use the same quota and billing rules as the hosted service.

The README includes configuration examples for Cursor, VS Code, Windsurf, Cline, Claude Code, and Claude Desktop through mcp-remote. Claude Desktop does not natively use HTTP, so it requires that adapter.

Tools and capabilities

• get_time_series: Retrieve weekly or daily history for one keyword on one source.
• get_growth: Compare percentage change across preset or custom time windows. Multiple windows for one source and keyword count as one request.
• get_top_trends: Retrieve current rankings, with optional limit, offset, category, sort, and change window parameters.
• Use source-specific keyword formats documented by the service.
• Request data from multiple sources with get_growth by passing a comma-separated source list.

Limitations and notes

News Sentiment Trends MCP MCP server requires one source per get_time_series call. Source names and keyword formats depend on the service catalog. For get_top_trends, the MCP type is mandatory and must match the feed name exactly; REST can omit it only when requesting every feed, which is billed per feed.

The free account tier provides 100 requests per month. The README does not define a complete list of supported sources or feed names, so consult the current service documentation for available catalogs and parameter details. Index values are normalized only where the relevant pipeline supports that behavior, and absolute volume is not available for every source.