Pippin

From side project to platform

Pippin started as a community-built documentation tool — the kind of thing a few engineers stand up because the sanctioned option doesn't fit how they actually work. By the time I took it on, the demand was already real; what it lacked was the reliability, integrations, and product discipline to be something people could depend on daily.

The work was equal parts platform engineering and product management: nothing here shipped without deciding what belonged in the product and what didn't.

Team and delivery

A major editor revamp shipped alongside accessibility features, real-time collaboration, spreadsheet support, agentic tools, and a review mode. None of it landed in isolation — it landed because the team could load-test for enterprise-scale reliability and had degradation instrumentation and fallback logic in place before traffic caught up with ambition.

Keeping up with a moving substrate

The hardest part of running an AI product internally isn't the product — it's that the tools underneath it keep changing out from under you. As usage patterns shifted (users migrating from one CLI tool to another, for instance), I kept feature parity intact through the Pippin MCP integration layer rather than chasing each migration with a rewrite.

I also led a community contribution framework that prioritized customer-impacting features and accelerated the convergence of what had been several competing internal tools into one. Consolidation is unglamorous work, and it's usually the difference between a platform and a pile of side projects.

Result

125,000+ monthly active users, 40,000+ daily active users, and 15–20% month-over-month growth sustained since January 2026 — for a tool that started as something a few engineers built because nothing else fit.