MCP Statistics

MCP ecosystem-wide metrics derived from data we are scraping from across the internet. The statistics on this page are kept dynamically up-to-date by our internal data warehouse, updated daily. They are meant to approximate the whole MCP ecosystem, and not the metrics of any single particular platform that integrates with MCP.

Specifically, what you see below are plots of an estimated downloads metric for MCP servers. This metric is:

• Our calculation as to how many times we think this (local) server was downloaded across the entire internet.
• Does not (yet) incorporate remote MCP servers.
• Not specific to PulseMCP in any way - it's an objective measure stretching across platforms and clients.
• Composed of a blend of registry download counters, social signals, web traffic, and more.
• Certainly not perfect. But it's the best unified metric for relative comparison anyone has published to-date, and we'll work to continue to improve it and keep it state-of-the-art for what it's meant to be.

We also display a "total MCP servers" metric in this data. Important notes about this count:

• This count is representative of PulseMCP's scraped database of meaningful MCP servers.
• We intentionally omit low quality implementations that we don't think would ever be used by someone besides the creator.
• This is why our "total count" may be less than that of another server aggregator; we are confident that the difference in total numbers can be purely attributed to our decision to not catalog the lowest quality implementations.

Like what you see? Want similar kinds of data? Shoot us a request at hello@pulsemcp.com and we'll see what we can do for you and the MCP community.