Pendo MCP

Pendo's MCP server lets an AI assistant query product analytics without opening the Pendo interface or building a report first. An agent can look up the applications, pages and features tracked in a subscription, pull usage over a chosen time range for any of them, and break the numbers down by visitor or account segment so the answer reflects the cohort the question was actually about. It also reaches the qualitative side of Pendo: in-app guides and how they performed, plus survey and NPS responses. That pairing is the interesting part, because an agent can line up a drop in feature usage against what users wrote in a survey over the same window and return both halves in a single answer instead of two separate exports. Access is scoped by the Pendo API key the server is configured with, so the agent sees one subscription and inherits whatever that key is permitted to read. The audience is product managers, product operations and customer success teams already instrumented with Pendo. Common uses are checking adoption after a launch, identifying accounts that never reached a key workflow, measuring whether a guide changed behavior, and assembling a business review without queueing work for an analyst.