Repolode
Submit
nashsu

nashsu/llm_wiki

Cross-platform desktop app that turns your documents into a persistent, interlinked wiki maintained by an LLM instead of re-querying from scratch.

Open-source alternative to:NotebookLM

View on GitHub
Stars
21,429
Forks
2,436
Open issues
286
Last commit
17 days ago

What is it?

LLM Wiki is a desktop application that ingests documents such as PDFs, Office files, EPUB/MOBI, web clips and URLs, then has an LLM build and maintain a wiki of interlinked Markdown pages. Knowledge is compiled once at ingest time and kept current as new sources arrive, rather than being re-derived on every query as in typical RAG setups. The wiki directory works as an Obsidian vault.

The app adds a knowledge graph with a relevance model, Louvain community detection, and graph insights that surface knowledge gaps and bridge pages. Chat runs through a Rust-backend agent with wiki, source, graph and web retrieval, and optional LanceDB vector search can be enabled per project. Deep Research uses Tavily, SerpApi or SearXNG to fill gaps and ingest the results.

It is built with Tauri v2, React and TypeScript, and provides a local HTTP API and MCP server on 127.0.0.1 so external agents can query projects. A Chrome extension clips web pages into the wiki. Licensed under GPL v3.

Key features

  • 01Two-step ingest where the LLM analyzes a source before generating wiki pages with source traceability.
  • 02Parses PDF, DOCX, PPTX, spreadsheets, EPUB/MOBI, images and web clips into Markdown.
  • 03Knowledge graph with relevance scoring, community detection and insights on knowledge gaps.
  • 04Optional vector search via LanceDB with any OpenAI-compatible embedding endpoint.
  • 05Persistent ingest queue with crash recovery, retry and cancel.
  • 06Local HTTP API and MCP server for external AI agents to query wiki content.

Use cases

  • Building a personal research knowledge base from papers, reports and web articles.
  • Keeping an Obsidian-compatible wiki up to date as new reading material arrives.
  • Letting a coding or chat agent query your local documents through the built-in API or MCP server.
  • Organizing study notes or project documentation into interlinked pages.

Things to know

  • The license could not be identified from the metadata, so usage terms are unclear.
  • Using the app requires configuring an LLM provider, and optional features like vector search need an embedding endpoint.

In collections

Desktop GUI apps & toolkits

Topics

#llm#knowledge-base#personal-knowledge-management#rag#desktop-app#tauri#knowledge-graph#obsidian

FAQ

What is llm_wiki?
Cross-platform desktop app that turns your documents into a persistent, interlinked wiki maintained by an LLM instead of re-querying from scratch. LLM Wiki is a desktop application that ingests documents such as PDFs, Office files, EPUB/MOBI, web clips and URLs, then has an LLM build and maintain a wiki of interlinked Markdown pages. Knowledge is compiled once at ingest time and kept current as new sources arrive, rather than being re-derived on every query as in typical RAG setups. The wiki directory works as an Obsidian vault.
Is llm_wiki free to use? What license does it have?
llm_wiki is open source under the Other license. Check the license terms before using it in commercial products.
Is llm_wiki still maintained?
Actively maintained. The last commit to llm_wiki was on Oct 21, 2026.
What can I use llm_wiki for?
Building a personal research knowledge base from papers, reports and web articles. Keeping an Obsidian-compatible wiki up to date as new reading material arrives. Letting a coding or chat agent query your local documents through the built-in API or MCP server. Organizing study notes or project documentation into interlinked pages.
Where can I get llm_wiki?
The source code and installation instructions are on GitHub: https://github.com/nashsu/llm_wiki

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects