wiki-langGraph: A LangGraph Pipeline for Building Obsidian-Style Markdown Wikis from Raw Sources
Summary
This article presents a LangGraph pipeline called "wiki-langGraph" that ingests raw markdown files and compiles them into an Obsidian-style wiki vault with wikilinks, backlinks, semantic "See also" links, and an Index.md. It optionally supports LLM-driven authoring, query-to-save workflows for compounding future context, local QMD index refreshing, and vault linting. The project is inspired by Andrej Karpathy's "LLM Wiki" pattern, where a model incrementally builds and maintains a persistent, interlinked markdown wiki rather than just retrieving chunks at query time (classic RAG). The pipeline is designed to help users structure and synthesize information from raw sources into a navigable knowledge base.
Source
Twitter / Xwiki-langGraph: A LangGraph Pipeline for Building Obsidian-Style Markdown Wikis from Raw Sourcesgithub.comKey quotes
· 3 pulledAndrej Karpathy posted on X about a workflow where the model does not only retrieve chunks at question time (classic RAG), but incrementally builds and maintains a persistent, interlinked markdown wiki between you and your raw sources—so structure and synthes
LangGraph pipeline that ingests raw markdown, compiles an Obsidian-style vault (wikilinks, backlinks, Index.md), optionally runs LLM authoring and semantic 'See also' links, refreshes a local QMD index when enabled, and lints the vault.
Background: the 'LLM Wiki' pattern
You might also wanna read
Evaluating LangGraph for Agentic AI Workflows: A Decision-Maker's Guide
LangGraph is becoming the default framework for teams building agentic AI workflows, but its growing reputation means many teams adopt it by
labyrinthanalyticsconsulting.com·1mo ago
Why plain markdown beats vector databases for feeding knowledge to LLMs
This essay argues that the AI industry has unnecessarily complicated knowledge delivery to LLMs using vector databases, SDKs, and proprietar
formaly.io·19h agoWhy plain markdown beats vector databases for feeding knowledge to LLMs
This essay argues that the AI industry has unnecessarily complicated knowledge delivery to LLMs using vector databases, SDKs, and proprietar
formaly.io·19h ago
Using Markdown as a Protocol for AI-Generated User Interfaces
The article explores the concept of using Markdown as a protocol for agentic AI interfaces, where AI assistants can generate reactive user i
fabian-kuebler.com·3mo ago
DokuWiki to Add Native Markdown Support in Upcoming Release
DokuWiki, the long-running open-source wiki software, is preparing to add native Markdown support in its next release. This will give users
linuxiac.com·13d ago
Proxy-Pointer RAG: Optimizing Knowledge Graph Ingestion by Reducing NER Overhead
This article discusses the Proxy-Pointer RAG architecture as a solution to the costly problem of Named Entity Recognition (NER) and relation
towardsdatascience.com·15d ago
Copy as Markdown: Tool Converts Web Content to Markdown Format for AI Language Models
The article introduces 'Copy as Markdown,' a tool that converts web content into clean Markdown format specifically optimized for use with L
Product Hunt·1y ago
Comments
Sign in to join the conversation.
No comments yet. Be the first.