TrendFlow
Live Google Trends research for AI.
TrendFlow is a hosted service that exposes Google Trends data over the Model Context Protocol and a REST API, so an AI assistant or your own code can query live search interest directly.
It makes Google Trends usable from wherever you already work. It speaks the Model Context Protocol over HTTP, so pointing ChatGPT, Claude, Perplexity or Cursor at the hosted endpoint gives the assistant live search-interest data. The same allowance is available as a plain REST API with bearer keys — any language, any backend, any cron job.
Interest over time, interest by region, trending now, related queries, search suggestions and the full geographic hierarchy are all covered, with type-safe enums for regions, timeframes and resolutions, and exports to JSON, CSV or a pandas DataFrame.
Two things make the data better than the obvious approach. Trending queries are read from Google’s internal RPC rather than the public RSS feed, which returns roughly five times as many items with real growth figures attached. And topic lookup resolves an entity rather than a literal string — a topic can score 62 where the exact phrase scores 1.
The libraries are open source and MIT licensed in both ecosystems, maintained at feature parity: trendflow-py for Python and trendflow for JavaScript and TypeScript, the latter with zero runtime dependencies. Trending searches and topic lookup are free forever, with no account.
How it compares
Against unofficial client libraries such as pytrends: you are trading a library you run yourself for a hosted endpoint someone else keeps working when Google changes something. The client libraries here are open source, but the service behind them is not self-hostable.
Against commercial trend products such as Exploding Topics or Glimpse: those blend multiple signals and editorialise what matters. TrendFlow does not. It is a faithful, typed interface to one source, which is the right choice when you want the underlying numbers and the wrong one when you want a curated answer.
It is Google Trends only. If you need social, marketplace or news signals alongside search interest, this is one input rather than the whole picture.
BUILT BY Dario Mory
STATUS Shipped — commercial
PRICING Free tier · Pro $12/mo · Business $99/mo
PLATFORM Hosted MCP server, REST API, Python and JavaScript libraries
STACK Next.js 15, Vercel, Neon Postgres, Model Context Protocol, Stripe
WEBSITE trendflow.mory.dev
Questions
What does TrendFlow cost?
There is a free tier, Pro at 12 USD per month and Business at 99 USD per month. Trending searches and topic lookup are free forever and need no account.
How do I connect Claude, ChatGPT or Cursor to it?
It speaks the Model Context Protocol over HTTP, so you point the assistant at the hosted MCP endpoint and it gets live search-interest data as a tool. No local server to run.
Can I use it without MCP?
Yes. The same allowance is available as a plain REST API with bearer keys, so any language, backend or cron job can call it.
What data can I get?
Interest over time, interest by region, trending now, related queries, search suggestions and the full geographic hierarchy, with type-safe enums for regions, timeframes and resolutions.
What export formats are supported?
JSON, CSV, or a pandas DataFrame directly from the Python library.
Why not just scrape Google Trends myself?
Two differences in the data. Trending queries are read from Google internal RPC rather than the public RSS feed, which returns roughly five times as many items with real growth figures attached. And topic lookup resolves an entity rather than a literal string, which matters more than it sounds: a topic can score 62 where the exact phrase scores 1.
Are the client libraries open source?
Yes, MIT licensed in both ecosystems and maintained at feature parity — trendflow-py for Python and trendflow for JavaScript and TypeScript, the latter with zero runtime dependencies.
