[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-top-ai-github-repositories-dominating-2026-zh":3,"article-related-top-ai-github-repositories-dominating-2026-zh":40,"series-tools-6f625844-2519-4fac-8be2-d2d06a0686f5":93},{"id":4,"title":5,"content":6,"summary":7,"source":8,"source_url":9,"author":10,"image_url":11,"keywords":12,"language":23,"translated_content":10,"views":24,"is_premium":25,"created_at":26,"updated_at":26,"cover_image":11,"published_at":27,"rewrite_status":28,"rewrite_error":10,"rewritten_from_id":29,"slug":30,"category":31,"related_article_id":32,"status":33,"google_indexed_at":10,"x_posted_at":10,"tweet_text":10,"title_rewritten_at":10,"title_original":10,"key_takeaways":34,"topic_cluster_id":38,"embedding":39,"is_canonical_seed":25},"6f625844-2519-4fac-8be2-d2d06a0686f5","2026 最強 AI GitHub 倉庫盤點","\u003Cp data-speakable=\"summary\">這篇在整理 \u003Ca href=\"\u002Fnews\u002Ftop-ai-prompt-engineering-tools-2026-zh\">2026\u003C\u002Fa> 年最受開發者關注的 AI \u003Ca href=\"\u002Fnews\u002Fhorizon-github-ai-news-briefings-zh\">GitH\u003C\u002Fa>ub 倉庫，重點放在能直接拿來做產品的框架、\u003Ca href=\"\u002Ftag\u002Fagent\">Agent\u003C\u002Fa>、在地模型與推論工具。\u003C\u002Fp>\u003Cp>說真的，\u003Ca href=\"\u002Fnews\u002Fgithub-skills-repos-turn-ai-coding-into-workflows-zh\">GitH\u003C\u002Fa>ub 還是 AI 開發的第一戰場。星星數很吵，但真正有用的是，哪些 repo 真的有人 clone、fork，還拿去上線。\u003C\u002Fp>\u003Cp>這篇不看虛的。看的是 \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fopenai\u002Fopenai-cookbook\" target=\"_blank\" rel=\"noopener\">OpenAI Cookbook\u003C\u002Fa>、\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Flangchain-ai\u002Flangchain\" target=\"_blank\" rel=\"noopener\">LangChain\u003C\u002Fa>、\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Frun-llama\u002Fllama_index\" target=\"_blank\" rel=\"noopener\">LlamaIndex\u003C\u002Fa>、\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fautogen\" target=\"_blank\" rel=\"noopener\">AutoGen\u003C\u002Fa> 這類專案。它們解的是實際問題，不是只會做 demo。\u003C\u002Fp>\u003Ctable>\u003Cthead>\u003Ctr>\u003Cth>Repository\u003C\u002Fth>\u003Cth>功能\u003C\u002Fth>\u003Cth>為什麼重要\u003C\u002Fth>\u003C\u002Ftr>\u003C\u002Fthead>\u003Ctbody>\u003Ctr>\u003Ctd>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fopenai\u002Fopenai-cookbook\" target=\"_blank\" rel=\"noopener\">OpenAI Cookbook\u003C\u002Fa>\u003C\u002Ftd>\u003Ctd>OpenAI API 範例\u003C\u002Ftd>\u003Ctd>直接給可抄的實作模式\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Flangchain-ai\u002Flangchain\" target=\"_blank\" rel=\"noopener\">LangChain\u003C\u002Fa>\u003C\u002Ftd>\u003Ctd>LLM 應用與 Agent 框架\u003C\u002Ftd>\u003Ctd>很多團隊的起手式\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Frun-llama\u002Fllama_index\" target=\"_blank\" rel=\"noopener\">LlamaIndex\u003C\u002Fa>\u003C\u002Ftd>\u003Ctd>資料與 LLM 整合框架\u003C\u002Ftd>\u003Ctd>文件檢索場景很常用\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fautogen\" target=\"_blank\" rel=\"noopener\">AutoGen\u003C\u002Fa>\u003C\u002Ftd>\u003Ctd>多 Agent 協作框架\u003C\u002Ftd>\u003Ctd>適合需要多步推理的工作流\u003C\u002Ftd>\u003C\u002Ftr>\u003C\u002Ftbody>\u003C\u002Ftable>\u003Ch2>星星數不是全部\u003C\u002Fh2>\u003Cp>很多人看 repo，第一眼只看 stars。老實說，這很偷懶。星星高，不代表你週五晚上能把功能生出來。\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1779136463477-0jml.png\" alt=\"2026 最強 AI GitHub 倉庫盤點\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>真正重要的是文件清不清楚、維護有沒有持續、範例能不能直接搬進專案。這些才決定一個 repo 是玩具，還是真工具。\u003C\u002Fp>\u003Cp>2026 年最值得看的 AI repo，通常都有一個共同點。它們幫你少走很多冤枉路。你不用先自己發明整套架構，才知道模型怎麼接資料、怎麼跑工具、怎麼串 Agent。\u003C\u002Fp>\u003Cul>\u003Cli>文件完整，導入成本低。\u003C\u002Fli>\u003Cli>版本更新穩定，代表還有人管。\u003C\u002Fli>\u003Cli>範例比宣傳圖更有用。\u003C\u002Fli>\u003Cli>API 乾淨，團隊比較好接。\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>開發者一直回頭用的幾個 repo\u003C\u002Fh2>\u003Cp>先講最常見的一組：\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fopenai\u002Fopenai-cookbook\" target=\"_blank\" rel=\"noopener\">OpenAI Cookbook\u003C\u002Fa>、\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Flangchain-ai\u002Flangchain\" target=\"_blank\" rel=\"noopener\">LangChain\u003C\u002Fa>、\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Frun-llama\u002Fllama_index\" target=\"_blank\" rel=\"noopener\">LlamaIndex\u003C\u002Fa>、\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fautogen\" target=\"_blank\" rel=\"noopener\">AutoGen\u003C\u002Fa>。這四個剛好涵蓋了 AI 應用最常見的四種需求：範例、編排、檢索、協作。\u003C\u002Fp>\u003Cp>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fopenai\u002Fopenai-cookbook\" target=\"_blank\" rel=\"noopener\">OpenAI Cookbook\u003C\u002Fa> 比較像實戰筆記。它不是完整框架，但你可以直接把範例改成產品程式。\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Flangchain-ai\u002Flangchain\" target=\"_blank\" rel=\"noopener\">LangChain\u003C\u002Fa> 的定位更廣，適合做鏈式流程和工具調用。\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Frun-llama\u002Fllama_index\" target=\"_blank\" rel=\"noopener\">LlamaIndex\u003C\u002Fa> 則更偏向資料接入，這點對企業案很重要。很多案子卡住的不是模型，而是資料爛得像倉庫。\u003C\u002Fp>\u003Cp>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fautogen\" target=\"_blank\" rel=\"noopener\">AutoGen\u003C\u002Fa> 走的是多 Agent 路線。講白了，就是讓幾個模型角色分工合作。這種做法不一定每個場景都需要，但一旦工作流很長，就很有感。\u003C\u002Fp>\u003Cblockquote>\"The next big thing in software will be systems that can reason, plan, and act,\" said Satya Nadella at Microsoft Build 2023.\u003C\u002Fblockquote>\u003Cp>這句話放到今天還是對味。AI 應用的重心，已經從單次問答，移到多步驟任務。\u003C\u002Fp>\u003Cp>你現在看到的 repo 熱度，其實就是這個轉向的縮影。大家不再只想要聊天機器人。大家想要的是能查資料、會呼叫工具、還能自己往下做的系統。\u003C\u002Fp>\u003Cul>\u003Cli>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Flangchain-ai\u002Flangchain\" target=\"_blank\" rel=\"noopener\">LangChain\u003C\u002Fa> 適合通用編排。\u003C\u002Fli>\u003Cli>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Frun-llama\u002Fllama_index\" target=\"_blank\" rel=\"noopener\">LlamaIndex\u003C\u002Fa> 適合檢索型產品。\u003C\u002Fli>\u003Cli>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fautogen\" target=\"_blank\" rel=\"noopener\">AutoGen\u003C\u002Fa> 適合多角色協作。\u003C\u002Fli>\u003Cli>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fopenai\u002Fopenai-cookbook\" target=\"_blank\" rel=\"noopener\">OpenAI Cookbook\u003C\u002Fa> 適合快速學 API 寫法。\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>在地模型和推論工具更實際\u003C\u002Fh2>\u003Cp>另一群很重要的 repo，是處理成本、延遲和隱私的工具。像 \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Follama\u002Follama\" target=\"_blank\" rel=\"noopener\">Ollama\u003C\u002Fa>、\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp\" target=\"_blank\" rel=\"noopener\">llama.cpp\u003C\u002Fa>、\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\" target=\"_blank\" rel=\"noopener\">vLLM\u003C\u002Fa>，都讓在地跑模型或高效率部署變簡單。\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1779136457309-v16s.png\" alt=\"2026 最強 AI GitHub 倉庫盤點\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>這件事很現實。很多團隊不是不想用 AI，而是怕費用炸掉，或資料不能亂出機房。\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Follama\u002Follama\" target=\"_blank\" rel=\"noopener\">Ollama\u003C\u002Fa> 讓本機實驗門檻很低。\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp\" target=\"_blank\" rel=\"noopener\">llama.cpp\u003C\u002Fa> 一直是高效率推論的代表。\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\" target=\"_blank\" rel=\"noopener\">vLLM\u003C\u002Fa> 則偏向高吞吐服務，適合真的有流量的產品。\u003C\u002Fp>\u003Cp>如果你在台灣做 B2B 軟體，這類工具更有感。客戶常常先問兩件事：成本多少，資料放哪裡。技術再酷，這兩題答不好，案子一樣卡。\u003C\u002Fp>\u003Cp>這裡也可以看出一個趨勢。模型 \u003Ca href=\"\u002Ftag\u002Fapi\">API\u003C\u002Fa> 不是唯一答案。很多團隊開始混用雲端 API、在地模型、快取和批次推論，目的很簡單，就是把錢省下來，把延遲壓下去。\u003C\u002Fp>\u003Cul>\u003Cli>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Follama\u002Follama\" target=\"_blank\" rel=\"noopener\">Ollama\u003C\u002Fa> 適合本機測試。\u003C\u002Fli>\u003Cli>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp\" target=\"_blank\" rel=\"noopener\">llama.cpp\u003C\u002Fa> 適合輕量推論。\u003C\u002Fli>\u003Cli>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\" target=\"_blank\" rel=\"noopener\">vLLM\u003C\u002Fa> 適合高流量服務。\u003C\u002Fli>\u003Cli>這三個都在壓低部署成本。\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>數字攤開來看更清楚\u003C\u002Fh2>\u003Cp>如果把這些 repo 放在一起比較，你會發現它們解的是不同層的問題。有人負責教你怎麼寫，有人負責讓你串資料，有人負責讓模型跑得快。\u003C\u002Fp>\u003Cp>下面這張表很直白。你可以直接拿來判斷，哪個 repo 比較適合你的情境。講白了，別把所有問題都丟給同一個框架。\u003C\u002Fp>\u003Cp>我自己會把它們分成三類。第一類是學習型。第二類是產品型。第三類是基礎設施型。這樣選起來比較不會亂。\u003C\u002Fp>\u003Ctable>\u003Cthead>\u003Ctr>\u003Cth>專案\u003C\u002Fth>\u003Cth>定位\u003C\u002Fth>\u003Cth>適合場景\u003C\u002Fth>\u003C\u002Ftr>\u003C\u002Fthead>\u003Ctbody>\u003Ctr>\u003Ctd>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fopenai\u002Fopenai-cookbook\" target=\"_blank\" rel=\"noopener\">OpenAI Cookbook\u003C\u002Fa>\u003C\u002Ftd>\u003Ctd>範例庫\u003C\u002Ftd>\u003Ctd>學 API、做 PoC\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Flangchain-ai\u002Flangchain\" target=\"_blank\" rel=\"noopener\">LangChain\u003C\u002Fa>\u003C\u002Ftd>\u003Ctd>通用框架\u003C\u002Ftd>\u003Ctd>Agent、工具調用、流程編排\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Frun-llama\u002Fllama_index\" target=\"_blank\" rel=\"noopener\">LlamaIndex\u003C\u002Fa>\u003C\u002Ftd>\u003Ctd>資料框架\u003C\u002Ftd>\u003Ctd>RAG、文件問答、內部知識庫\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fautogen\" target=\"_blank\" rel=\"noopener\">AutoGen\u003C\u002Fa>\u003C\u002Ftd>\u003Ctd>多 Agent 框架\u003C\u002Ftd>\u003Ctd>複雜任務拆解、協作式工作流\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Follama\u002Follama\" target=\"_blank\" rel=\"noopener\">Ollama\u003C\u002Fa>\u003C\u002Ftd>\u003Ctd>在地推論\u003C\u002Ftd>\u003Ctd>本機測試、內網部署\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fllama.cpp\" target=\"_blank\" rel=\"noopener\">llama.cpp\u003C\u002Fa>\u003C\u002Ftd>\u003Ctd>高效率推論\u003C\u002Ftd>\u003Ctd>輕量硬體、邊緣裝置\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\" target=\"_blank\" rel=\"noopener\">vLLM\u003C\u002Fa>\u003C\u002Ftd>\u003Ctd>服務端推論\u003C\u002Ftd>\u003Ctd>高吞吐、低延遲服務\u003C\u002Ftd>\u003C\u002Ftr>\u003C\u002Ftbody>\u003C\u002Ftable>\u003Cp>這份比較也反映 2026 年的開發現實。AI 開發已經不是單點工具。它更像一條供應鏈。前面有模型，中間有框架，後面有推論和監控。\u003C\u002Fp>\u003Cp>你如果只學框架，不懂推論，最後還是會被成本打臉。你如果只玩模型，不懂資料，產品一樣答非所問。這就是現在最常見的坑。\u003C\u002Fp>\u003Ch2>這些 repo 背後的產業脈絡\u003C\u002Fh2>\u003Cp>AI 開源社群這幾年很明顯，重心一直在往「可上線」移動。以前大家愛看模型參數，現在大家更在意延遲、成本、評估和可維護性。\u003C\u002Fp>\u003Cp>這也解釋了為什麼框架型專案和推論型專案會一起紅。前者解決開發速度，後者解決營運壓力。兩邊缺一個，產品都不完整。\u003C\u002Fp>\u003Cp>如果再往前看，這種變化其實很像雲端時代。早期大家比誰先做出功能，後來就開始比誰撐得住流量、誰的架構比較好維護。AI 現在也走到這一步了。\u003C\u002Fp>\u003Cp>我覺得接下來最重要的不是再多一個聊天 demo，而是更好的 eval、觀測和部署工具。畢竟模型再聰明，出錯時還是要有人收尾。\u003C\u002Fp>\u003Ch2>接下來該怎麼選\u003C\u002Fh2>\u003Cp>如果你現在要開始，我的建議很簡單。先選一個框架 repo，再選一個推論 repo。不要一開始就全都碰，會很亂。\u003C\u002Fp>\u003Cp>最實際的組合是 \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Flangchain-ai\u002Flangchain\" target=\"_blank\" rel=\"noopener\">LangChain\u003C\u002Fa> 或 \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Frun-llama\u002Fllama_index\" target=\"_blank\" rel=\"noopener\">LlamaIndex\u003C\u002Fa>，再搭配 \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Follama\u002Follama\" target=\"_blank\" rel=\"noopener\">Ollama\u003C\u002Fa> 或 \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\" target=\"_blank\" rel=\"noopener\">vLLM\u003C\u002Fa>。這樣你可以先把產品跑起來，再慢慢補 Agent、評估和監控。\u003C\u002Fp>\u003Cp>我的預測很直接。2026 年會更吃重「能不能穩定上線」，而不是「星星有多少」。如果你是開發者，現在最值得做的事，就是挑一個 repo，真的做一個可用功能出來。\u003C\u002Fp>","整理 2026 年最受開發者關注的 AI GitHub 倉庫，包含框架、Agent、在地模型與推論工具，幫你快速看懂哪些專案真的能上線。","medium.com","https:\u002F\u002Fmedium.com\u002Flets-code-future\u002Ftop-ai-github-repositories-that-are-dominating-2026-and-you-also-need-to-know-about-that-970ce6a61b96",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1779136463477-0jml.png",[13,14,15,16,17,18,19,20,21,22],"AI GitHub repo","LangChain","LlamaIndex","AutoGen","Ollama","llama.cpp","vLLM","OpenAI Cookbook","AI 工具","Agent 框架","zh",0,false,"2026-05-18T20:33:51.029803+00:00","2026-05-18T20:33:50.82+00:00","done","b8f22776-3899-4e07-a146-f567d72f2626","top-ai-github-repositories-dominating-2026-zh","tools","6dad42df-02b8-437e-ac11-6f687dec68be","published",[35,36,37],"星星數只能當參考，文件、維護和實作範例更重要。","LangChain、LlamaIndex、AutoGen 各自解不同層的問題。","Ollama、llama.cpp、vLLM 代表 2026 年更重視在地推論與部署效率。","c3c88dd2-a940-438a-b359-0e5a24562273","[-0.0371364,-0.0026290533,0.013333965,-0.06856854,-0.012158544,-0.004601817,-0.027029352,0.0009761861,0.006268726,-0.0023156335,-0.0063027674,-0.020133292,0.016612653,0.0063275453,0.11847426,0.0018754584,-0.00796529,0.021805143,0.03595052,-0.039224997,0.014586121,-0.0058819694,0.026383257,0.0007260464,-0.0057596937,-0.004189362,-0.0029078675,0.0028811193,0.057154216,0.014988094,-0.002915247,0.0025981,-0.003216606,-0.0019790675,0.0039840727,0.022917729,0.015689641,0.0026857755,0.033589296,-0.0052250223,-0.00914199,0.0025400151,-0.00044692642,-0.009556006,-0.014556692,0.0075404206,-0.003927794,-0.032278717,-0.0067898207,0.0037685267,-0.022108678,0.015128407,-0.020714257,-0.14745289,0.0008479482,0.012741801,-0.0137194935,0.021628225,0.016199335,-0.0038231276,-0.004869223,0.01630762,-0.030903043,-0.00239037,0.0011509656,-0.013076157,0.023623928,0.0022125917,-0.0052806255,-0.031017343,-0.034562428,-0.007230917,0.0010458677,-0.023215676,0.030385757,-0.003542654,-0.0097403135,0.008480026,0.014900575,0.024902258,0.008140131,-0.021598864,0.016257279,0.0037260843,-0.016321404,-0.012738545,-0.0046706977,0.0031724088,0.0053890175,0.017950915,0.014251644,0.03171678,-0.0040457253,0.0030878393,-0.0071308855,0.0034651507,0.0024739176,-0.0019952932,0.010733319,-0.012718047,0.0075550387,-0.026651317,0.003913922,-3.9749397e-05,0.020588279,-0.008022152,0.026332427,0.007661132,-0.021773543,-0.007588803,0.01528319,-0.040085614,0.005953634,0.019588478,0.020585934,-0.122804284,-0.01120502,-0.007826464,0.019785002,0.0169948,-0.0060387035,0.00766554,-0.0029520101,-0.0052138115,-0.0029005455,-0.011945472,0.017904796,0.010613418,-0.0038511981,0.0056395847,-0.010833561,0.021933114,-0.012247763,-0.01594313,0.004326662,0.031193592,-0.015098903,-0.011401517,0.0040766993,-0.003507675,-0.008005214,0.020114858,-0.004847405,-0.0026387086,-0.049227588,0.0028912656,-0.028233027,-0.000674065,-0.00029582917,-0.015641192,-0.0008201546,-0.018945673,-0.0016861934,-0.012548689,0.0030599916,-0.00038557642,0.0026479298,0.009332779,0.014231515,0.016860235,-0.011325327,-0.009306264,-0.0061157485,0.030078877,0.016493024,0.025310256,-0.015789915,0.015120468,0.01688719,0.016157689,0.007771505,0.0055257953,0.0043026265,0.0024556566,0.0021339837,0.009469572,0.019992607,-0.0045827935,0.006529063,-0.005433165,-0.010339176,0.00563319,0.002057633,0.00078227173,0.0031921272,-0.02567871,-0.009923104,0.012216251,0.0011391925,0.004367743,-0.027327513,-0.0067559127,0.0176534,-0.021465398,-0.0031799164,-0.011505291,0.007570817,0.0014467379,-0.027514659,0.0131169055,0.009969685,-0.0036325387,0.022152027,-0.025192462,-0.00018523459,-0.00795285,-0.039284125,-0.021542123,-0.013922683,-0.026794408,-0.034144893,0.009145011,0.026253259,-0.025990002,-0.023463437,-0.012694624,-0.012375982,-0.005955088,0.02569778,0.0010042249,-0.007887143,-0.014573877,0.0040978207,0.023158355,-0.002526043,-0.013304944,0.0051566726,-0.020125693,-0.05552271,0.0030346424,0.014458618,0.006914345,-0.019246833,0.013892041,-0.0011228608,-0.00017835794,-0.008694057,0.020831602,0.015732022,0.021706115,-0.016169183,0.045602128,0.0030968508,0.0007325519,0.018758493,-0.032437358,0.00012353282,0.017175227,0.002444248,0.012379856,-0.014617911,-0.011072963,-0.008464895,0.0235988,-0.005049685,0.02811887,-0.0040472965,0.014241647,-0.01710812,0.013862423,-0.01299746,-0.0036093504,-0.005116978,-0.0076796073,0.012810736,-0.005324872,0.022396272,0.0055684177,-0.026939968,0.02692708,0.00042965915,0.020233395,0.0134032825,-0.019802678,0.025024515,-0.0027397233,-0.059603926,0.023723837,-0.0028145572,-0.023549253,0.002132673,0.020062437,-0.015750285,0.0017070739,-0.011011814,0.007138262,-0.00825051,-0.010785964,0.0010124508,-0.0058500324,0.008975204,0.020696636,-0.012680949,-0.012827913,0.0018154791,-0.018329179,-0.0037986622,0.010643278,-0.017172435,-0.0026235718,0.019015342,-0.008491262,0.009256965,0.042463526,0.0014309316,0.011911498,0.014978965,0.030041624,0.017696735,0.0041730576,-0.003469591,-0.019662919,0.012462394,-0.035104416,-0.009813102,-0.025354424,0.00042378565,-0.032369204,-0.011956101,-0.0042191464,-0.025122738,-0.017631914,-0.010907097,-0.0077438084,-0.029102212,0.0047804355,0.002520208,0.017789079,0.021561505,0.0019420817,0.015512561,-4.994635e-05,0.029915085,-0.010861551,-0.002838345,0.010996344,0.0015226133,-0.0043046544,-0.03436398,0.004255595,-0.0041357228,-0.012576814,-0.017678937,0.017103741,-0.008701998,0.0132131735,0.019270789,-0.0057664416,-0.023557357,-0.026809502,0.044511124,-0.0086693065,-0.005224334,-0.0075517627,-0.035932664,-0.011250454,-0.021859279,0.010083777,0.023850424,0.0014402752,0.014285746,-0.026319575,-0.0021313715,-0.020566786,0.02951241,-0.011447108,0.013352013,0.02147645,-0.02585021,0.007271015,-0.009595839,0.01663898,0.010938043,-0.014399575,0.01357348,0.010298571,-0.009184182,-0.0017429262,-0.021848278,-0.0019369462,-0.0016622324,0.0045876345,-0.037973337,0.0017222803,-0.013720291,-0.019706994,0.037844546,0.0095679555,0.014612407,-0.038178608,-0.0027368863,-0.012994819,-0.010249433,0.004368978,-0.00048580553,-0.008563474,-0.0221215,0.014757817,-0.02210815,-0.0060069403,0.036127403,-0.003946922,-0.0038879812,0.015454752,-0.005767287,0.0068065533,0.0032511582,0.025280675,0.0049050734,-0.01689234,0.0075038723,0.019124376,0.010159444,0.014638007,0.022568624,0.021501416,0.0025274262,-0.005676678,-0.007349844,-0.023927357,-0.03371894,0.022766173,0.008924595,0.020054435,-0.012134731,0.008757073,-0.025752496,0.022075495,-0.00051456847,0.0044553853,-0.028317856,0.014906762,-0.012842391,-0.03421833,-0.029534774,-0.0018382197,-0.01369296,-0.014328268,-0.03240324,-0.0059758457,0.012354492,0.0011061414,0.024387825,0.011314129,0.0031853542,-0.027572451,0.0105074225,-0.01890949,-0.0419608,0.015277424,0.024386983,0.0024832543,0.018627647,-0.009889614,-0.012855616,-0.006645728,-0.019971183,0.024986386,-0.008747226,-0.025479753,-0.0096956445,-0.0008605954,-0.012721157,0.028429452,0.014922508,-0.031752788,0.0013128243,-0.0038312427,-0.0027013212,0.014616966,-0.022304857,-0.016075995,0.026986474,0.0008715359,-0.010178903,-0.00457593,0.011822744,0.017238496,0.01977013,-0.0017175263,-0.000678849,-0.011466883,0.025973322,-0.0022658927,0.018151563,0.025219355,0.00020988681,-0.005819278,-0.0061957627,0.017437423,0.012833803,0.019154815,0.043953404,0.0012303076,-0.03357159,-0.02393415,-0.013481042,-0.0049984604,0.012143575,0.0071573774,-0.003642455,-0.03090757,-0.0014452052,0.007198037,0.004104683,-0.01872383,0.0019242666,0.00544267,0.0072003314,0.005507801,0.0075428686,-0.028513499,0.0055700713,0.014120233,-0.007818843,-0.025505748,-0.015637744,-0.0018849099,0.0026630242,-0.009270816,-0.015901033,-0.02331484,-0.020935412,0.016385753,-0.0023393023,-0.036036674,0.0342863,0.017609881,0.005263066,0.018940173,-0.009899059,0.01901207,-0.017874138,0.020570409,0.015558222,-0.004547466,0.014671666,0.035817593,-0.014533187,0.004665545,-0.02376921,-0.024109168,-0.000252894,-0.024413025,0.046380926,-0.08945988,-0.02605288,0.012614691,-0.027951218,0.0053974064,-0.0068375426,-0.0049478244,-0.020419616,-0.0027619281,0.029362181,0.023173612,0.011018799,0.004544883,0.0020687093,-0.0063118003,-0.0036820746,0.0049288142,-0.026731407,0.038553413,-0.041204482,0.015578557,-0.021288661,0.015091043,-0.0056751883,-0.0065685506,-0.008610485,0.02713042,-0.009162412,-0.01959461,-0.0050029214,-0.039254934,-0.018490639,-0.004360247,0.010525855,-0.016459571,0.01624674,0.012310991,-0.010479651,-0.0068560895,-0.0032417772,0.020765288,0.0056522456,-0.029045312,-0.004380157,-0.007439435,0.010358432,0.011708482,-0.007807499,0.00775338,0.025335303,-0.042117164,-0.016898,-0.011349319,-0.03943548,0.00602016,-0.0051244595,-0.0022027886,0.0109701995,-0.0037271015,0.0030435603,-0.022631802,-0.008040682,-0.016865062,0.03144983,-0.021471225,0.03320495,0.011213823,0.024271687,0.0003869628,-0.015225381,0.0014407192,-0.014812333,0.022527536,0.03721094,-0.011326099,0.027382137,-0.000636889,0.005107336,-0.0047151754,-0.01878999,-0.013246765,0.0021875757,-0.1025085,-0.02320358,-0.024420414,-0.0049320017,-0.007897093,-0.0041468283,0.0017679355,-0.017318983,-0.010851646,-0.043485254,0.004033494,0.004274364,-0.022195335,-0.03123807,0.030288985,0.0041802265,0.0033550982,0.008707177,0.0031633696,-0.012248034,-0.02130659,0.0054270006,0.03735129,-0.029434528,-0.0031169828,0.03227994,0.0023802775,0.018964874,0.008582338,-0.012533998,-0.026740605,-0.14869243,-0.027577708,-0.016555117,0.0017081157,0.012486572,0.035439342,0.0017599694,0.014859291,0.0019959384,-0.018755214,-0.0082858065,-0.017407743,-0.018005144,0.0041277683,-0.016333802,0.111240245,0.012285139,0.028067932,-0.009263171,-0.018729314,0.008787433,-0.0077274693,-0.04145694,0.0134781245,0.00328346,-0.00428955,0.023496468,-0.013023643,-0.014984596,0.02441203,0.009919931,0.003001094,-0.010021084,-0.025659258,0.015012764,0.008722246,-0.00036553814,-0.027303614,0.044491746,0.010680745,0.004861663,0.026298415,0.025516873,-0.024820527,-0.026730116,0.008947827,-0.014330449,-0.027896767,-0.025398655,0.0069565643,-0.050292205,-0.049899403,-0.00716859,-0.026972,0.02094805,0.008851797,-0.023628369,-0.0047210604,0.029103497,0.020464525,0.0310657,-0.0060297674,-0.005370942,0.03636901,0.0028470259,0.024109729,0.013909555,0.029550713,0.0005883436,0.021735698,0.007871975,0.0054023936,-0.003748592,-0.023301939,-0.023387954,-0.037104536,-0.002205924,0.04272715,0.038385082,-0.004998874,0.007206646,-0.00030899208,-0.0037466236,0.006657355,0.043081824,0.016107908,-0.0031985561,-0.015212763,0.014213285,0.0028352484,0.012834149,0.03310501,-0.01152953,-0.0073586125,0.014023382,0.021218449,-0.007310281,0.02773382,0.015441041,0.01332216,0.004568019,-0.02959041,0.039271817,-0.014695667,0.017495926,0.017462336,0.0061009713,0.018032843,0.02006583,0.0051504793]",{"tags":41,"relatedLang":52,"relatedPosts":56},[42,44,46,48,50],{"name":13,"slug":43},"ai-github-repo",{"name":14,"slug":45},"langchain",{"name":17,"slug":47},"ollama",{"name":16,"slug":49},"autogen",{"name":15,"slug":51},"llamaindex",{"id":32,"slug":53,"title":54,"language":55},"top-ai-github-repositories-dominating-2026-en","Top AI GitHub Repositories Dominating 2026","en",[57,63,69,75,81,87],{"id":58,"slug":59,"title":60,"cover_image":61,"image_url":61,"created_at":62,"category":31},"eb8ba5dd-3093-4c8e-8625-bfd5f11f3800","zhihu-soft-copywriting-four-tips-zh","知乎軟文教你寫出能賣貨的標題","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1779146044192-aild.png","2026-05-18T23:13:42.179515+00:00",{"id":64,"slug":65,"title":66,"cover_image":67,"image_url":67,"created_at":68,"category":31},"8c6ebfe4-617a-4380-a06b-69b5a80bdde2","github-copilot-app-desktop-agent-preview-zh","GitHub Copilot 桌面代理預覽開放","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1779141829436-mot7.png","2026-05-18T22:03:22.898714+00:00",{"id":70,"slug":71,"title":72,"cover_image":73,"image_url":73,"created_at":74,"category":31},"1e852f64-41ab-476c-83a6-d0c8732f86b1","xurrent-q2-2026-ai-release-turns-itsm-into-agents-zh","Xurrent Q2 AI 讓 ITSM 變成代理","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1779140063861-o0jh.png","2026-05-18T21:33:56.788859+00:00",{"id":76,"slug":77,"title":78,"cover_image":79,"image_url":79,"created_at":80,"category":31},"41688e9f-54e8-4270-b887-769090f99622","horizon-github-ai-news-briefings-zh","Horizon：GitHub AI 新聞簡報庫","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1779135836565-0x4g.png","2026-05-18T20:23:28.561989+00:00",{"id":82,"slug":83,"title":84,"cover_image":85,"image_url":85,"created_at":86,"category":31},"4ecd2dc5-dc13-4f9b-86bd-0a97f2f814d7","github-skills-repos-turn-ai-coding-into-workflows-zh","GitHub 技能庫把 AI 寫程式變流程","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1779134707718-sslr.png","2026-05-18T20:04:26.488566+00:00",{"id":88,"slug":89,"title":90,"cover_image":91,"image_url":91,"created_at":92,"category":31},"7d325b37-35ff-4ecb-b9d4-a6cab8d0c48e","top-ai-prompt-engineering-tools-2026-zh","2026 必看 AI Prompt 工具整理","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1779133443961-ijso.png","2026-05-18T19:43:34.399846+00:00",[94,99,104,109,114,119,124,129,134,139],{"id":95,"slug":96,"title":97,"created_at":98},"de769291-4574-4c46-a76d-772bd99e6ec9","googles-biggest-gemini-launches-in-2026-zh","Google 2026 最大 Gemini 盤點","2026-03-26T07:26:39.21072+00:00",{"id":100,"slug":101,"title":102,"created_at":103},"855cd52f-6fab-46cc-a7c1-42195e8a0de4","surepath-real-time-mcp-policy-controls-zh","SurePath 推出即時 MCP 政策控管","2026-03-26T07:57:40.77233+00:00",{"id":105,"slug":106,"title":107,"created_at":108},"9b19ab54-edef-4dbd-9ce4-a51e4bae4ebb","mcp-in-2026-the-ai-tool-layer-teams-use-zh","2026 年 MCP：團隊真的在用的 AI 工具層","2026-03-26T08:01:46.589694+00:00",{"id":110,"slug":111,"title":112,"created_at":113},"af9c46c3-7a28-410b-9f04-32b3de30a68c","prompting-in-2026-what-actually-works-zh","2026 提示工程，真正有用的是什麼","2026-03-26T08:08:12.453028+00:00",{"id":115,"slug":116,"title":117,"created_at":118},"05553086-6ed0-4758-81fd-6cab24b575e0","garry-tan-open-sources-claude-code-toolkit-zh","Garry Tan 開源 Claude Code 工具包","2026-03-26T08:26:20.068737+00:00",{"id":120,"slug":121,"title":122,"created_at":123},"042a73a2-18a2-433d-9e8f-9802b9559aac","github-ai-projects-to-watch-in-2026-zh","2026 必看 20 個 GitHub AI 專案","2026-03-26T08:28:09.619964+00:00",{"id":125,"slug":126,"title":127,"created_at":128},"a5f94120-ac0d-4483-9a8b-63590071ac6a","claude-code-vs-cursor-2026-zh","Claude Code 與 Cursor 深度對比：202…","2026-03-26T13:27:14.279193+00:00",{"id":130,"slug":131,"title":132,"created_at":133},"0975afa1-e0c7-4130-a20d-d890eaed995e","practical-github-guide-learning-ml-2026-zh","2026 機器學習入門 GitHub 實用指南","2026-03-27T01:16:49.712576+00:00",{"id":135,"slug":136,"title":137,"created_at":138},"bfdb467a-290f-4a80-b3a9-6f081afb6dff","aiml-2026-student-ai-ml-lab-repo-review-zh","AIML-2026：像課綱的學生實驗 Repo","2026-03-27T01:21:51.467798+00:00",{"id":140,"slug":141,"title":142,"created_at":143},"80cabc3e-09fc-4ff5-8f07-b8d68f5ae545","ai-trending-github-repos-and-research-feeds-zh","AI Trending：把 AI 資源收成一張表","2026-03-27T01:31:35.262183+00:00"]