PiSharp

Problem

AVEVA PI System is everywhere in industrial operations, but its developer ecosystem is thin and scattered: PI Web API answers live in PDFs, vendor forums, and tribal knowledge, with no vendor-independent, implementation-first home. PI engineers lose hours hunting for working code, authentication patterns, and integration guidance that should be one search away.

Approach

We built PiSharp as a single Next.js 16 / React 19 application on a Drizzle + Neon Postgres backend, structured as a stack of businesses — Knowledge, then Workflow Software, Community, and Services. The Developer Hub ships a PI Web API track of 20+ first-hand guides (start-here, authentication, reading/writing values, batch requests, WebID lookup, pagination limits, SDK comparisons, troubleshooting) plus a downloadable Cookbook of 12 production-ready Python recipes and a Postman collection. An AI Q&A assistant (Vercel AI SDK, with session history and usage limits) answers PI questions in context; a community forum carries the discussion with automated moderation via Gemini 2.5 Flash against published guidelines; and a jobs board with resume parsing rounds out the platform. Auth runs on Better Auth; analytics on PostHog.

Outcome

PiSharp is live at pisharp.com as Basira's own product, with the Developer Hub, AI assistant, community, and packaged PI services (integration audit, quickstart, data-pipeline sprint) shipped and in active iteration. It demonstrates, end to end, our ability to build and operate a real content + AI + community product in a deep industrial niche — and it's the home of our open-source PI Web API Python SDK.