Documenting a Power Platform solution shouldn't mean expanding every node of a 200-action flow. 😅
I built pp-pedia, a Power Platform documentation generator and AI assistant that runs 100% in your browser.
Drop in your solution ZIP and it will:
📦 Extract and parse the solution (flows, apps, tables, columns, choices, web resources)
📝 Generate readable documentation, with an interactive reader
🔎 Build vector embeddings in a Web Worker and store everything locally in PGlite
📤 Export the documentation when you need it
Then ask it questions: (optional, bring your own key)
💬 "What's the blast radius if I rename the cr_approval_status column?"
💬 "Which flows use premium connectors?"
💬 "If I make cr_tax_id required, which flows or apps might fail on record creation?"
💬 "Do any of our JS web resources use deprecated APIs like Xrm.Page?"
Why client-side?
Solution archives often contain business logic and schema details you can't upload to a random server. Here nothing leaves your machine. The only exception is if you choose to plug in an LLM provider, and then only the queries and retrieved context go to that provider.
Under the hood: a RAG pipeline with an agentic loop that calls client-side tools, so answers are grounded in your actual solution rather than guesses.
It's open source (Apache-2.0), so try it, break it, and tell me what's missing. 🙌
🔗 Try it: https://sagarpandey88.github.io/pp-pedia/
⭐ Code: https://github.com/sagarpandey88/pp-pedia
Top comments (0)