dheerapat

dheerapat / pi-kb

Public

A pi extension that compiles documents into a structured, interlinked wiki using your LLM

agent knowledge-base pi
15
0
89% credibility
Found May 20, 2026 at 15 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

pi-kb is a knowledge base extension for a coding assistant that helps you build a personal wiki from your documents. You can add markdown files from your computer or fetch pages from the web using simple commands. The AI reads each document, creates summaries, and identifies topics that connect across your documents. Everything is stored as plain markdown files in a folder on your computer, making it easy to search, back up, or open in other tools. You can ask questions against your compiled knowledge base and get answers that draw from all your documents.

How It Works

1
💡 You want to build a personal wiki

You've been collecting research papers, notes, and articles and wish you had one place to search and explore them all.

2
🔌 You install the knowledge base extension

You add the extension to your coding assistant, giving it the ability to organize and link your documents.

3
📄 You add your first document

You type a simple command to add a markdown file or paste a web URL, and the tool saves it to your knowledge base.

4
🤖 Your AI reads and understands the document

The assistant reads your document, writes a summary of the key ideas, and spots topics that connect to your other documents.

5
Your wiki grows in different ways
📝
Adding files from your computer

You point to markdown files on your machine and they get copied into your wiki.

🌐
Adding pages from the web

You paste a URL and the tool fetches the page, converts it to markdown, and saves it.

6
🔍 You ask your knowledge base questions

You type a question and the assistant searches your compiled wiki to find and synthesize the relevant information.

🎉 You have a living, searchable wiki

Your documents are organized, linked, and ready to explore. Everything lives as plain files you can open anywhere, even in tools like Obsidian.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 15 to 15 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is pi-kb?

pi-kb is a TypeScript extension for the pi coding agent that turns your markdown documents into a searchable, interlinked wiki. Drop in a README, a blog post, or a URL, and the extension uses your LLM to extract summaries, identify cross-cutting concepts, and stitch everything into a graph-ready knowledge base stored as plain markdown files. The wiki lives at `~/.pi/agent/kb/` and works with Obsidian's graph view out of the box.

Why is it gaining traction?

Most knowledge base tools require you to manually organize and tag content. pi-kb offloads that work to your LLM -- it reads the wiki state, writes summaries, merges overlapping topics into concepts, and maintains the index, all via simple slash commands like `/kb-add` and `/kb-query`. It also fetches and converts web pages to markdown automatically, so building a technical reference wiki is nearly frictionless. The plain-file storage approach means your knowledge base is portable, version-controlled with Git, and accessible without any special tooling.

Who should use this?

Developers who work across many documentation sources -- API specs, architecture decision records, blog posts, research -- and want a unified, queryable reference. Researchers and technical writers building domain-specific knowledge graphs will also benefit. If you're already using pi as your coding agent and want persistent, context-aware memory across sessions, this fills that gap without requiring external services.

Verdict

The core idea is solid and the implementation is surprisingly complete for a small project, but the credibility score of 0.9% and 15 stars mean this is early-stage software with limited real-world validation. Try it if you're already in the pi ecosystem and want a lightweight wiki layer; otherwise, wait for a more mature release or community feedback before betting a workflow on it.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.