[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-llm-wiki-compiler-raw-sources-to-wiki-en":3,"article-related-llm-wiki-compiler-raw-sources-to-wiki-en":33,"series-industry-f50df767-8d6e-4ceb-886f-f8cb610a9f19":86},{"id":4,"slug":5,"title":6,"content":7,"summary":8,"source":9,"source_url":10,"author":11,"image_url":12,"cover_image":12,"category":13,"language":14,"translated_content":11,"related_article_id":15,"keywords":16,"key_takeaways":25,"views":29,"created_at":30,"published_at":31,"topic_cluster_id":32},"f50df767-8d6e-4ceb-886f-f8cb610a9f19","llm-wiki-compiler-raw-sources-to-wiki-en","llm-wiki-compiler turns raw sources into a wiki","\u003Cp data-speakable=\"summary\">llm-wiki-compiler turns raw sources into an interlinked, cited wiki.\u003C\u002Fp>\n\u003Cp>GitHub’s \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fatomicstrata\u002Fllm-wiki-compiler\">atomicstrata\u002Fllm-wiki-compiler\u003C\u002Fa> packages a two-phase pipeline, a local viewer, and agent-friendly export paths for durable knowledge work. The repo has about 1.5k stars and 155 forks.\u003C\u002Fp>\n\u003Ctable>\u003Cthead>\u003Ctr>\u003Cth>Item\u003C\u002Fth>\u003Cth>Primary payoff\u003C\u002Fth>\u003Cth>Notable detail\u003C\u002Fth>\u003C\u002Ftr>\u003C\u002Fthead>\u003Ctbody>\u003Ctr>\u003Ctd>Compiled wiki\u003C\u002Ftd>\u003Ctd>Structured output\u003C\u002Ftd>\u003Ctd>Typed pages with citations\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Hybrid retrieval\u003C\u002Ftd>\u003Ctd>Better evidence selection\u003C\u002Ftd>\u003Ctd>Embeddings, BM25, graph expansion\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Local viewer\u003C\u002Ftd>\u003Ctd>Browse and inspect\u003C\u002Ftd>\u003Ctd>Search, graph, citation chips\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Eval harness\u003C\u002Ftd>\u003Ctd>Quality checks\u003C\u002Ftd>\u003Ctd>Health score and regression deltas\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>MCP server\u003C\u002Ftd>\u003Ctd>Agent access\u003C\u002Ftd>\u003Ctd>Context packs for Claude Desktop and Cursor\u003C\u002Ftd>\u003C\u002Ftr>\u003C\u002Ftbody>\u003C\u002Ftable>\n\u003Ch2>1. Compiled wiki output\u003C\u002Fh2>\n\u003Cp>The core promise is simple: feed in raw material and get back a persistent wiki instead of a pile of loose chunks. The compiler turns sources into typed pages such as concept, entity, comparison, and overview, with paragraph- and claim-level citations tied back to source line ranges.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781389970201-6yvr.png\" alt=\"llm-wiki-compiler turns raw sources into a wiki\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\n\u003Cp>That matters when you want a reference artifact that can be read later without re-running the whole discovery step. The repo is inspired by Karpathy’s LLM Wiki pattern, but it adds explicit provenance and page typing so the output is easier to trust and reuse.\u003C\u002Fp>\n\u003Cul>\n  \u003Cli>Inputs: notes, docs, papers, READMEs, ADRs\u003C\u002Fli>\n  \u003Cli>Outputs: interlinked markdown pages\u003C\u002Fli>\n  \u003Cli>Citations: source line ranges at claim level\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>2. Hybrid retrieval pipeline\u003C\u002Fh2>\n\u003Cp>Search is not just vector lookup. The project uses incremental, content-hash-aware embeddings to narrow the candidate set, then BM25 reranking and wikilink-graph expansion to assemble the final evidence pack.\u003C\u002Fp>\n\u003Cp>That mix helps when a source base grows past a few files and recall starts to matter more than raw similarity. It also means the compiler can keep retrieval focused while still surfacing connected pages that a pure embedding pass might miss.\u003C\u002Fp>\n\u003Cul>\n  \u003Cli>Semantic chunk embeddings for top-K narrowing\u003C\u002Fli>\n  \u003Cli>BM25 reranking for lexical precision\u003C\u002Fli>\n  \u003Cli>Graph expansion for linked context\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>3. Local web viewer\u003C\u002Fh2>\n\u003Cp>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fatomicstrata\u002Fllm-wiki-compiler\">llmwiki view\u003C\u002Fa> opens a read-only browser UI for inspecting the compiled wiki. You get sidebar navigation, search, a force-directed graph, and provenance chips on each page, which makes it easier to trace where a claim came from.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781389969773-r763.png\" alt=\"llm-wiki-compiler turns raw sources into a wiki\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\n\u003Cp>This is the part that turns the compiled output into something people can actually audit. Instead of reading raw logs or JSON, you browse the knowledge base the way you would browse a small internal encyclopedia.\u003C\u002Fp>\n\u003Cul>\n  \u003Cli>Read-only browser interface\u003C\u002Fli>\n  \u003Cli>Sidebar navigation and search\u003C\u002Fli>\n  \u003Cli>Graph view plus citation chips\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>4. Eval harness and health checks\u003C\u002Fh2>\n\u003Cp>The repo includes \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fatomicstrata\u002Fllm-wiki-compiler\">llmwiki eval\u003C\u002Fa>, which scores wiki health from 0 to 100 and reports citation coverage, precision, and regression deltas. It can also use LLM-as-judge support scoring and threshold checks that fit into CI.\u003C\u002Fp>\n\u003Cp>For teams, this is a practical guardrail. You can tell whether a new ingest improved the wiki, broke citations, or quietly degraded quality before the changes spread into downstream use.\u003C\u002Fp>\n\u003Ccode>llmwiki eval --threshold 85 --judge --json\u003C\u002Fcode>\n\u003Ch2>5. Freshness, rollback, and audit history\u003C\u002Fh2>\n\u003Cp>Knowledge bases rot when sources move. This project tracks stale and orphaned pages, supports targeted repair with \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fatomicstrata\u002Fllm-wiki-compiler\">llmwiki refresh --stale\u003C\u002Fa>, and writes a durable operation log so each ingest, compile, and query is recorded.\u003C\u002Fp>\n\u003Cp>It also adds rollback and diff-oriented reports, which is useful when you need to reverse a bad ingest or explain why a page changed. The repo even caches the latest lint summary in .llmwiki\u002Flast-lint.json so viewer health can show recent results without rerunning lint.\u003C\u002Fp>\n\u003Cul>\n  \u003Cli>Stale-claim checks and freshness reports\u003C\u002Fli>\n  \u003Cli>Reverse ingest and compile diff reports\u003C\u002Fli>\n  \u003Cli>Timestamped log.md audit trail\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>6. MCP server and in-process SDK\u003C\u002Fh2>\n\u003Cp>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fatomicstrata\u002Fllm-wiki-compiler\">llmwiki serve\u003C\u002Fa> exposes the pipeline through MCP, so tools like Claude Desktop, \u003Ca href=\"\u002Ftag\u002Fcursor\">Cursor\u003C\u002Fa>, and \u003Ca href=\"\u002Fnews\u002Ffable-5-claude-code-like-coworker-en\">Claude Code\u003C\u002Fa> can ask for budgeted, citation-aware context packs. That makes the compiler usable as an agent memory layer instead of only a standalone CLI.\u003C\u002Fp>\n\u003Cp>For developers who want direct integration, createWiki({ root }) runs ingest, compile, query, status, freshness, export, and eval inside your own process. That is a cleaner fit for custom tooling than shelling out on every step.\u003C\u002Fp>\n\u003Ccode>createWiki({ root }).query(\"what changed?\")\u003C\u002Fcode>\n\u003Ch2>7. Provider support and export paths\u003C\u002Fh2>\n\u003Cp>The tool is built to work across multiple model backends, including \u003Ca href=\"\u002Ftag\u002Fanthropic\">Anthropic\u003C\u002Fa>, the \u003Ca href=\"\u002Ftag\u002Fclaude\">Claude\u003C\u002Fa> Agent SDK, \u003Ca href=\"\u002Ftag\u002Fopenai\">OpenAI\u003C\u002Fa>-compatible servers, Ollama, and \u003Ca href=\"\u002Ftag\u002Fgithub-copilot\">GitHub Copilot\u003C\u002Fa>. It also exports typed JSON envelopes that can be imported into \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fatomicmemory\u002Fllmwiki\">@atomicmemory\u002Fllmwiki\u003C\u002Fa> as verbatim Atomic Memory records.\u003C\u002Fp>\n\u003Cp>That portability matters if your team mixes hosted APIs with local models. You can keep the same workflow while changing only the provider settings, which lowers the cost of trying the project in real environments.\u003C\u002Fp>\n\u003Cul>\n  \u003Cli>Anthropic and Claude Agent SDK support\u003C\u002Fli>\n  \u003Cli>OpenAI-compatible local servers and Ollama\u003C\u002Fli>\n  \u003Cli>JSON export for runtime memory systems\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>How to decide\u003C\u002Fh2>\n\u003Cp>Pick this project if you want a source-to-wiki pipeline that keeps citations, supports agents, and leaves an audit trail. It fits researchers, technical writers, and maintainers who need durable reference material rather than a one-off summary.\u003C\u002Fp>\n\u003Cp>If your main need is quick search over a few files, the full compiler may be more than you need. If you want a browsable knowledge base that can be refreshed, evaluated, and handed to agents, this repo is built for that job.\u003C\u002Fp>","8 features show how llm-wiki-compiler turns notes, docs, and papers into a linked wiki with citations and audit trails.","github.com","https:\u002F\u002Fgithub.com\u002Fatomicstrata\u002Fllm-wiki-compiler",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781389970201-6yvr.png","industry","en","4ea87c5d-e6f5-4b0b-8725-886c8bf90724",[17,18,19,20,21,22,23,24],"llm wiki compiler","knowledge compiler","interlinked wiki","MCP server","citation-aware retrieval","agent memory","source freshness","audit trail",[26,27,28],"It compiles raw sources into typed wiki pages with claim-level citations.","It includes retrieval, viewer, eval, freshness, rollback, and audit features.","It supports CLI, SDK, MCP, and multiple model providers.",0,"2026-06-13T22:32:23.002929+00:00","2026-06-13T22:32:22.993+00:00","5fe38f8a-dc8c-44bd-a3dc-82024f24ba0f",{"tags":34,"relatedLang":45,"relatedPosts":49},[35,37,39,41,43],{"name":20,"slug":36},"mcp-server",{"name":21,"slug":38},"citation-aware-retrieval",{"name":40,"slug":40},"llm-wiki-compiler",{"name":19,"slug":42},"interlinked-wiki",{"name":18,"slug":44},"knowledge-compiler",{"id":15,"slug":46,"title":47,"language":48},"llm-wiki-compiler-raw-sources-to-wiki-zh","7 個把原始資料編成可追溯 Wiki 的功能","zh",[50,56,62,68,74,80],{"id":51,"slug":52,"title":53,"cover_image":54,"image_url":54,"created_at":55,"category":13},"d5852d0a-88c2-4d93-869f-03e83e76424c","jensen-huang-lg-ai-cooperation-five-bets-en","Jensen Huang’s LG deal spans five AI bets","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781399872771-tzny.png","2026-06-14T01:17:22.547908+00:00",{"id":57,"slug":58,"title":59,"cover_image":60,"image_url":60,"created_at":61,"category":13},"59e63043-a1a2-4dab-a053-d64991e3f1cf","nvidia-sk-group-expand-ai-ties-co-development-en","Nvidia and SK Group expand AI ties into co-development","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781398976022-afrl.png","2026-06-14T01:02:30.330777+00:00",{"id":63,"slug":64,"title":65,"cover_image":66,"image_url":66,"created_at":67,"category":13},"59cc7315-b242-44f4-96ea-4adce886cf9e","pythons-jit-future-hangs-on-new-pep-en","Python’s JIT future hangs on a new PEP","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781389066269-7pur.png","2026-06-13T22:17:20.722886+00:00",{"id":69,"slug":70,"title":71,"cover_image":72,"image_url":72,"created_at":73,"category":13},"0dc4656a-15f8-48ec-bcdf-f9aca8ac32db","ukraines-ai-war-network-faster-combat-en","Ukraine’s AI war network points to faster combat","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781363881146-5og8.png","2026-06-13T15:17:22.601823+00:00",{"id":75,"slug":76,"title":77,"cover_image":78,"image_url":78,"created_at":79,"category":13},"49f998be-b115-4206-977f-40ca930d85ce","anthropic-governance-market-story-en","Anthropic’s governance debate is now a market story","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781362977290-fdxs.png","2026-06-13T15:02:18.113309+00:00",{"id":81,"slug":82,"title":83,"cover_image":84,"image_url":84,"created_at":85,"category":13},"109408c4-f132-4827-a4a9-0332b51a0bb7","mastercard-ai-payments-solana-bull-case-en","Mastercard’s AI Payments Move Is A Solana Bull Case, Not A Hype Story","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1781360267532-n3ga.png","2026-06-13T14:17:21.37458+00:00",[87,92,97,102,107,112,117,122,127,132],{"id":88,"slug":89,"title":90,"created_at":91},"d35a1bd9-e709-412e-a2df-392df1dc572a","ai-impact-2026-developments-market-en","AI's Impact in 2026: Key Developments and Market Shifts","2026-03-25T16:20:33.205823+00:00",{"id":93,"slug":94,"title":95,"created_at":96},"5ed27921-5fd6-492e-8c59-78393bf37710","trumps-ai-legislative-framework-en","Trump's AI Legislative Framework: What's Inside?","2026-03-25T16:22:20.005325+00:00",{"id":98,"slug":99,"title":100,"created_at":101},"e454a642-f03c-4794-b185-5f651aebbaca","nvidia-gtc-2026-key-highlights-innovations-en","NVIDIA GTC 2026: Key Highlights and Innovations","2026-03-25T16:22:47.882615+00:00",{"id":103,"slug":104,"title":105,"created_at":106},"0ebb5b16-774a-4922-945d-5f2ce1df5a6d","claude-usage-diversifies-learning-curves-en","Claude Usage Diversifies, Learning Curves Emerge","2026-03-25T16:25:50.770376+00:00",{"id":108,"slug":109,"title":110,"created_at":111},"69934e86-2fc5-4280-8223-7b917a48ace8","openclaw-ai-commoditization-concerns-en","OpenClaw's Rise Raises Concerns of AI Model Commoditization","2026-03-25T16:26:30.582047+00:00",{"id":113,"slug":114,"title":115,"created_at":116},"b4b2575b-2ac8-46b2-b90e-ab1d7c060797","google-gemini-ai-rollout-2026-en","Google's Gemini AI Rollout Extended to 2026","2026-03-25T16:28:14.808842+00:00",{"id":118,"slug":119,"title":120,"created_at":121},"6e18bc65-42ae-4ad0-b564-67d7f66b979e","meta-llama4-fabricated-results-scandal-en","Meta's Llama 4 Scandal: Fabricated AI Test Results Unveiled","2026-03-25T16:29:15.482836+00:00",{"id":123,"slug":124,"title":125,"created_at":126},"bf888e9d-08be-4f47-996c-7b24b5ab3500","accenture-mistral-ai-deployment-en","Accenture and Mistral AI Team Up for AI Deployment","2026-03-25T16:31:01.894655+00:00",{"id":128,"slug":129,"title":130,"created_at":131},"5382b536-fad2-49c6-ac85-9eb2bae49f35","mistral-ai-high-stakes-2026-en","Mistral AI: Facing High Stakes in 2026","2026-03-25T16:31:39.941974+00:00",{"id":133,"slug":134,"title":135,"created_at":136},"9da3d2d6-b669-4971-ba1d-17fdb3548ed5","cursors-meteoric-rise-pressures-en","Cursor's Meteoric Rise Faces Industry Pressures","2026-03-25T16:32:21.899217+00:00"]