satish860

satish860 / llm-kb

Public

LLM-powered knowledge base. Drop documents, build a wiki, ask questions. Inspired by Karpathy.

16
2
100% credibility
Found Apr 07, 2026 at 16 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

llm-kb turns a folder of documents like PDFs and spreadsheets into an interactive chat-based knowledge base that builds and improves a concept-organized wiki over time.

How It Works

1
📖 Discover the tool

You hear about a simple way to turn your document folder into a smart helper that answers questions from your files.

2
🛠️ Get it ready

You add the tool to your computer in one easy step and connect it to a thinking service for smart answers.

3
📁 Choose your folder

Point the tool at the folder holding your PDFs, spreadsheets, or notes – it finds everything inside.

4
🚀 Launch and prepare

Hit start, and it reads your files, builds a summary, and opens a chat window ready for your questions.

5
💬 Ask away

Type a question like 'What are the key points?' and get a clear answer with references to your exact pages.

6
Add more anytime

Drop new files into the folder while chatting, and it automatically updates everything without stopping.

Smarter every time

Your knowledge helper builds a growing wiki of insights, answering faster and better with each question you ask.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 16 to 16 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 llm-kb?

llm-kb is a TypeScript CLI tool for turning a folder of documents into an LLM-powered knowledge base. Drop PDFs, docs, spreadsheets, or images into a directory, run `llm-kb run ./docs`, and it parses files, builds a searchable index, and opens an interactive TUI chat powered by Anthropic models via Pi SDK or API key. Repeated queries update a concept-organized wiki, making answers faster and cached for future use—perfect for building LLM-powered knowledge bases on GitHub without complex setup.

Why is it gaining traction?

It stands out with a self-improving data flywheel: every question refines the wiki, boosting hit rates from 0% to 66% after dozens of queries, while file watchers auto-reindex new drops. CLI commands like `llm-kb query "summarize revenue"` or `llm-kb eval` deliver quick insights or quality checks with reports on gaps and waste, skipping the boilerplate of custom RAG pipelines. Developers dig the persistent sessions and TUI that mimics Claude's web UI, blending ease with power for LLM KB experimentation.

Who should use this?

Legal researchers sifting case PDFs, financial analysts querying quarterly reports, or security teams extracting cyber threat intelligence into LLM-powered knowledge graphs. Ideal for solo devs prototyping LLM-powered applications or enterprise folks needing quick KBQA over mixed docs without cloud dependencies.

Verdict

Early-stage gem at 16 stars and 1.0% credibility—solid README, 38 tests, MIT license, but watch for edge cases in parsing. Grab it if you need instant local LLM KB; skip for production-scale until more battle-tested.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.