[{"data":1,"prerenderedAt":4},["ShallowReactive",2],{"readme:llm_wiki":3},"\u003Ch1>LLM Wiki\u003C\u002Fh1>\n\u003Cp align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fnashsu\u002Fllm_wiki\u002FHEAD\u002Flogo.jpg\" width=\"128\" height=\"128\" alt=\"LLM Wiki Logo\" \u002F>\n\u003C\u002Fp>\u003Cp align=\"center\">\n  \u003Cstrong>A personal knowledge base that builds itself.\u003C\u002Fstrong>\u003Cbr \u002F>\n  LLM reads your documents, builds a structured wiki, and keeps it current.\n\u003C\u002Fp>\u003Cp align=\"center\">\n  \u003Ca href=\"#what-is-this\" rel=\"nofollow ugc noopener\">What is this?\u003C\u002Fa> •\n  \u003Ca href=\"#what-we-changed--added\" rel=\"nofollow ugc noopener\">Features\u003C\u002Fa> •\n  \u003Ca href=\"#tech-stack\" rel=\"nofollow ugc noopener\">Tech Stack\u003C\u002Fa> •\n  \u003Ca href=\"#installation\" rel=\"nofollow ugc noopener\">Installation\u003C\u002Fa> •\n  \u003Ca href=\"#credits\" rel=\"nofollow ugc noopener\">Credits\u003C\u002Fa> •\n  \u003Ca href=\"#license\" rel=\"nofollow ugc noopener\">License\u003C\u002Fa>\n\u003C\u002Fp>\u003Cp align=\"center\">\n  English | \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fnashsu\u002Fllm_wiki\u002Fblob\u002FHEAD\u002FREADME_CN.md\" rel=\"nofollow ugc noopener\">中文\u003C\u002Fa> | \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fnashsu\u002Fllm_wiki\u002Fblob\u002FHEAD\u002FREADME_JA.md\" rel=\"nofollow ugc noopener\">日本語\u003C\u002Fa> | \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fnashsu\u002Fllm_wiki\u002Fblob\u002FHEAD\u002FREADME_KO.md\" rel=\"nofollow ugc noopener\">한국어\u003C\u002Fa>\n\u003C\u002Fp>\u003Chr \u002F>\n\u003Cp align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fnashsu\u002Fllm_wiki\u002FHEAD\u002Fassets\u002Foverview.jpg\" width=\"100%\" alt=\"Overview\" \u002F>\n\u003C\u002Fp>\u003Ch2>Features\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Cstrong>Two-Step Chain-of-Thought Ingest\u003C\u002Fstrong> — LLM analyzes first, then generates wiki pages with source traceability and incremental cache\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Multimodal Image Ingestion\u003C\u002Fstrong> — extract embedded images from PDFs, generate factual captions with a vision LLM, surface them in image-aware search results with lightbox preview and jump-to-source\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Multi-format Document Parsing\u003C\u002Fstrong> — ingest PDF, Office documents, EPUB\u002FMOBI, Org mode, images, media, web clips, and batches of URLs, with built-in, cloud, or local MinerU PDF processing\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Flexible Model Configuration\u003C\u002Fstrong> — configure models per project, route Chat and Ingest independently, and manage custom providers, headers, and streaming output\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Source-grounded Retrieval\u003C\u002Fstrong> — use Read Sources Only mode to answer exclusively from original imported material\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Project Management & Migration\u003C\u002Fstrong> — export and import complete project archives across devices, and rebuild the Wiki index from existing pages\u003C\u002Fli>\n\u003Cli>\u003Cstrong>4-Signal Knowledge Graph\u003C\u002Fstrong> — relevance model with direct links, source overlap, Adamic-Adar, and type affinity\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Louvain Community Detection\u003C\u002Fstrong> — automatic knowledge cluster discovery with cohesion scoring\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Graph Insights\u003C\u002Fstrong> — surprising connections and knowledge gaps with one-click Deep Research\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Vector Semantic Search\u003C\u002Fstrong> — optional embedding-based retrieval via LanceDB, supports any OpenAI-compatible endpoint\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Persistent Ingest Queue\u003C\u002Fstrong> — serial processing with crash recovery, cancel, retry, and progress visualization\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Folder Import\u003C\u002Fstrong> — recursive folder import preserving directory structure, folder context as LLM classification hint\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Source Folder Auto-Watch\u003C\u002Fstrong> — detects external changes in \u003Ccode>raw\u002Fsources\u002F\u003C\u002Fcode> and keeps ingest\u002Fdelete cleanup in sync\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Deep Research\u003C\u002Fstrong> — LLM-optimized search topics, multi-query web search via Tavily, SerpApi, or SearXNG, auto-ingest results into wiki\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Rust Backend Chat Agent\u003C\u002Fstrong> — tool-using chat runtime with wiki\u002Fsource\u002Fgraph\u002Fweb retrieval, workspace file generation, shell approval, cancellation, and streaming tool events\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Agent Skills\u003C\u002Fstrong> — scan and enable local \u003Ccode>SKILL.md\u003C\u002Fcode> folders, select skills with \u003Ccode>\u002Fskill\u003C\u002Fcode>, and let the Agent read skill instructions on demand\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Generated Outputs Preview\u003C\u002Fstrong> — Agent-created Markdown, HTML, images, and other workspace files appear as outputs with preview and quick folder access\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Mermaid Diagram Rendering\u003C\u002Fstrong> — render Mermaid code blocks directly in chat and preview, with compact syntax-error cards instead of raw parser output\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Async Review System\u003C\u002Fstrong> — LLM flags items for human judgment, predefined actions, pre-generated search queries\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Chrome Web Clipper\u003C\u002Fstrong> — one-click web page capture with auto-ingest into knowledge base\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Local HTTP API + MCP Server + AI Agent Skill\u003C\u002Fstrong> — built-in \u003Ccode>127.0.0.1:19828\u003C\u002Fcode> JSON API and bundled MCP server for hybrid search, file read, graph traversal, and source rescan; ready-made \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fnashsu\u002Fllm_wiki_skill\" rel=\"nofollow ugc noopener\">agent skill\u003C\u002Fa> installs into Claude Code \u002F Codex with one command (\u003Ccode>npx skills add …\u003C\u002Fcode>)\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>What is this?\u003C\u002Fh2>\n\u003Cp>LLM Wiki is a cross-platform desktop application that turns your documents into an organized, interlinked knowledge base — automatically. Instead of traditional \u003C\u002Fp>\n",1790587590714]