Trends MCP for Claude

Trends MCP connects Claude to live trend data from 15+ sources: Google Search, Google Shopping, TikTok, Reddit, YouTube, Amazon, Wikipedia, news sentiment and volume, web traffic, app downloads, Steam, npm, and PyPI. After a one-time setup, any Claude conversation can ask questions like "what is accelerating on TikTok this week?" or "compare Amazon and Google Shopping demand for this product" and get answers built on current data instead of training-cutoff guesses.

Why Claude needs a live data connection

Claude's training data has a cutoff, and the cutoff is exactly where trend questions begin. Search interest in Claude AI itself shows why this matters: Trends MCP data pulled on August 5, 2026 puts Google Search interest in Claude AI up 500% year over year, from a normalized 12 to 72, even after cooling 15.29% from a July peak of 85. The ecosystem around Claude is moving just as fast. The official mcp Python package reached 88.8 million weekly PyPI downloads, up 1654.39% year over year from 5.1 million and still up 26.1% in the last 30 days. The npm SDK @modelcontextprotocol/sdk sits at 23.0 million weekly downloads, up 332.67% year over year from 5.3 million.

A user base growing that fast asks questions about a world that changes weekly. Without a live data connection, Claude answers trend questions from memory. With Trends MCP connected, Claude answers from this week's data and can cite the exact dates and figures behind every claim.

Setup on Claude.ai (browser)

The browser path takes under a minute and needs no config file:

1. Open Claude.ai Settings, then Connectors.
2. Choose Add custom connector and paste the Trends MCP server URL from the dashboard.
3. Complete the OAuth prompt in the browser.
4. Start a new chat and ask a trend question.

Claude.ai Pro, Max, and Team plans support remote MCP connectors. No API key handling is required on this path because authentication runs through OAuth.

Setup on Claude Desktop

The desktop path uses the standard MCP config file:

1. Copy the API key from the Trends MCP dashboard.
2. Open claude_desktop_config.json (Mac: ~/Library/Application Support/Claude/claude_desktop_config.json, Windows: %APPDATA%\Claude\claude_desktop_config.json).
3. Add the Trends MCP server under mcpServers with the server URL and the key in the headers.
4. Fully quit and restart Claude Desktop.

After restart, the Trends MCP tools appear in the tools list and Claude calls them automatically when a question needs trend data.

Setup on Claude Code

Claude Code supports MCP servers in its own configuration. Register the same remote server URL and API key, and the tools become available inside coding sessions. This is the path developers use when trend data feeds into programmatic work, for example the pipeline patterns in how to build a trends dashboard.

What to ask once connected

Time series analysis: "Chart Google Search interest in ai agents over the last year and explain the shape." Claude calls get_time_series(keyword='ai agents', source='google search') and narrates the result.

Cross-platform growth: "Compare mushroom coffee demand across Amazon, Google Shopping, TikTok, and Google Search over 1, 3, and 12 months." One get_growth call with a comma-separated source list returns all four platforms; the e-commerce research page shows what that workflow looks like for product teams.

Live boards: "What is trending on TikTok and Reddit right now?" Claude pulls get_top_trends for each of the 21 available feed types requested and summarizes what is breaking today.

The free tier includes 100 requests per month, enough for weekly research sessions. Paid plans start at $19/month for 1,000 requests, with Pro at $49/month for 5,000 requests. For a feature comparison against browser-based trend tools, see Trends MCP vs Glimpse.