[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-1932":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":9,"languages":9,"totalLinesOfCode":9,"stars":10,"forks":11,"watchers":12,"openIssues":13,"contributorsCount":13,"subscribersCount":13,"size":13,"stars1d":14,"stars7d":15,"stars30d":16,"stars90d":13,"forks30d":13,"starsTrendScore":17,"compositeScore":18,"rankGlobal":9,"rankLanguage":9,"license":19,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":22,"hasPages":20,"topics":23,"createdAt":9,"pushedAt":9,"updatedAt":24,"readmeContent":25,"aiSummary":26,"trendingCount":13,"starSnapshotCount":13,"syncStatus":27,"lastSyncTime":28,"discoverSource":29},1932,"llm-wiki","tonbistudio\u002Fllm-wiki","tonbistudio","Open-source template for building LLM-powered knowledge bases following Karpathy's LLM Wiki pattern",null,182,55,4,0,11,19,38,33,5.24,"MIT License",false,"master",true,[],"2026-06-12 02:00:34","# LLM Wiki\n\nAn open-source template for building LLM-powered knowledge bases, following [Andrej Karpathy's \"LLM Wiki\" pattern](https:\u002F\u002Fgist.github.com\u002Fkarpathy\u002F442a6bf555914893e9891c11519de94f).\n\nYou provide raw sources. The LLM reads them, writes structured wiki pages, cross-links everything, and maintains it over time. You never edit the wiki directly — you curate sources and ask questions.\n\n## How It Works\n\nThe system has three layers:\n\n```\nraw\u002F              Sources you collect (articles, transcripts, notes, PDFs)\nwiki\u002F             LLM-written & maintained pages (summaries, concepts, entities, syntheses)\nCLAUDE.md         Schema that tells the LLM how to structure everything\n```\n\nThree operations drive the workflow:\n\n| Operation | Trigger | What happens |\n|-----------|---------|--------------|\n| **Ingest** | \"ingest raw\u002Fmy-source.txt\" | LLM reads the source, creates a summary page, creates\u002Fupdates concept and entity pages, adds cross-links, updates the index and log |\n| **Query** | Ask any question | LLM searches the wiki, synthesizes an answer with citations, optionally creates a synthesis page for novel insights |\n| **Lint** | \"lint\" or \"health check\" | LLM audits all pages for orphans, contradictions, missing links, incomplete sections, and low-confidence claims — fixes what it can, reports the rest |\n\n## Quick Start\n\n1. **Clone this repo**\n   ```bash\n   git clone https:\u002F\u002Fgithub.com\u002FYOUR_USERNAME\u002Fllm-wiki.git my-knowledge-base\n   cd my-knowledge-base\n   ```\n\n2. **Customize CLAUDE.md** for your domain\n   - Update the Purpose section with your topic\n   - Replace the placeholder tagging taxonomy with your own categories\n   - Adjust confidence level descriptions if needed\n   - Everything else (workflows, page formats, linking rules) works as-is\n\n3. **Drop sources into `raw\u002F`**\n   - Text files, transcripts, articles, notes — any plain text\n   - These are immutable once added; the LLM never modifies them\n\n4. **Tell the LLM to ingest**\n   ```\n   ingest raw\u002Fmy-first-source.txt\n   ```\n   The LLM will create summary pages, concept pages, entity pages, cross-links, and update the index.\n\n5. **Ask questions**\n   ```\n   What are the key differences between X and Y?\n   ```\n   The LLM answers from the wiki, citing specific pages.\n\n6. **Run health checks**\n   ```\n   lint\n   ```\n   The LLM audits the wiki and fixes issues.\n\n## Directory Structure\n\n```\n.\n├── CLAUDE.md                      # Schema — the LLM's instructions\n├── raw\u002F                           # Your source documents (immutable)\n└── wiki\u002F\n    ├── index.md                   # Master catalog of all pages\n    ├── log.md                     # Append-only activity log\n    ├── dashboard.md               # Dataview dashboard (Obsidian)\n    ├── analytics.md               # Charts View analytics (Obsidian)\n    ├── flashcards.md              # Spaced repetition cards\n    ├── summaries\u002F                 # One page per source document\n    ├── concepts\u002F                  # Concept and framework pages\n    ├── entities\u002F                  # People, tools, organizations, etc.\n    ├── syntheses\u002F                 # Cross-cutting analyses and comparisons\n    ├── journal\u002F                   # Research\u002Fsession journal entries\n    │   └── template.md            # Journal entry template\n    └── presentations\u002F             # Marp slide decks\n```\n\n## Enhancements\n\nThis template includes several extras beyond the core wiki pattern:\n\n### Dataview Dashboard (`wiki\u002Fdashboard.md`)\nLive queries that surface low-confidence pages, recent updates, concepts by tag, and pages with the most sources. Requires the [Dataview](https:\u002F\u002Fgithub.com\u002Fblacksmithgu\u002Fobsidian-dataview) Obsidian plugin.\n\n### Charts View Analytics (`wiki\u002Fanalytics.md`)\nVisual analytics with pie charts, bar charts, and word clouds. Requires the [Charts View](https:\u002F\u002Fgithub.com\u002Fcaronchen\u002Fobsidian-chartsview-plugin) Obsidian plugin.\n\n### Mermaid Diagrams\nUse Mermaid code blocks in any wiki page to create flowcharts, sequence diagrams, or concept maps. Native support in Obsidian and GitHub.\n\n### Marp Slides (`wiki\u002Fpresentations\u002F`)\nCreate slide decks from markdown using [Marp](https:\u002F\u002Fmarp.app\u002F). Drop presentation files in this directory.\n\n### Research Journal (`wiki\u002Fjournal\u002F`)\nTrack your research sessions, experiments, or applied work with the included template. The LLM can reference journal entries when answering queries.\n\n### Spaced Repetition (`wiki\u002Fflashcards.md`)\nFlashcards in the format used by the [Spaced Repetition](https:\u002F\u002Fgithub.com\u002Fst3v3nmw\u002Fobsidian-spaced-repetition) Obsidian plugin. Ask the LLM to generate flashcards from any wiki page.\n\n### MCP Server\nThis repo works with Claude Code's MCP server capabilities. Point an MCP-compatible client at this repo and the LLM can read\u002Fwrite the wiki programmatically.\n\n## Customizing for Your Domain\n\nThe schema in `CLAUDE.md` is domain-agnostic. To adapt it:\n\n1. **Purpose** — Describe your knowledge domain in one paragraph\n2. **Tagging taxonomy** — Replace placeholder categories with your own (e.g., for a cooking KB: `cuisine`, `technique`, `ingredient`, `equipment`)\n3. **Confidence levels** — Adjust the descriptions to match your domain's evidence standards\n4. **Entity types** — Update the entity page description to match what entities mean in your domain (people, tools, companies, etc.)\n5. **Journal template** — Customize `wiki\u002Fjournal\u002Ftemplate.md` for your workflow\n\nEverything else — page format, linking conventions, workflows, rules — is universal and works across domains.\n\n## Example Domains\n\nThis template works for any knowledge-intensive topic:\n\n- **Research notes** — papers, experiments, methodologies\n- **Book analysis** — themes, characters, author techniques\n- **Competitive analysis** — companies, products, market trends\n- **Course notes** — lectures, readings, key concepts\n- **Personal development** — frameworks, habits, book summaries\n- **Technical documentation** — APIs, architectures, design patterns\n- **Hobby deep-dives** — any subject you want to master\n\n## License\n\nMIT\n","LLM Wiki 是一个开源模板，用于构建基于大语言模型的知识库，遵循 Andrej Karpathy 提出的 LLM Wiki 模式。其核心功能包括自动从原始资料生成结构化的维基页面、交叉链接以及持续维护知识库内容。用户只需提供原始资料并提出问题，无需直接编辑维基页面。该系统分为三个层次：原始资料层、由LLM创建和维护的维基页面层，以及定义结构模式的CLAUDE.md文件。通过“ingest”、“query”和“lint”三种操作驱动工作流，分别负责处理新资料、回答问题及检查与修复知识库中的问题。适用于需要高效管理和利用大量文本信息的场景，如学术研究、企业内部知识管理等。",2,"2026-06-11 02:46:52","CREATED_QUERY"]