[{"data":1,"prerenderedAt":222},["ShallowReactive",2],{"site-settings":3,"$fZv_QyiFkS5z18PFcvVRUOTTTXXmQH2gsiom1IxT8vBU":26,"related-posts":50,"$fxY7hrSdMOdy0-vFRUo48VbGytOZcKY3qDq4idD2qmOw":221},{"id":4,"site_name":5,"site_subtitle":6,"site_description":7,"author":8,"email":9,"github":10,"icp":10,"page_size":11,"enable_comment":12,"enable_rss":12,"updated_at":13,"site_icon":14,"site_og_image":15,"site_keywords":16,"site_og_image_width":24,"site_og_image_height":25},1,"LiteSite（轻站）","把你想做的网站说出来","LiteSite（轻站）探索自然语言驱动的零代码建站体验，展示响应式网站与前端原型的设计方向。","Admin","support@litesite.com",null,10,true,"2026-10-05T03:56:44.502+00:00","https:\u002F\u002Fnklbhnvyoggxghtaoizu.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-images\u002F2026-10-05-1765bce7-01f1-4206-8d5d-6a717d17f206.png","https:\u002F\u002Fnklbhnvyoggxghtaoizu.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-images\u002F2026-10-05-157c3c1d-2722-474c-860b-d3f38d6a674d.png",[17,18,19,20,21,22,23],"LiteSite","轻站","零代码建站","AI建站","响应式网站","前端原型","网站设计",1200,630,{"id":27,"slug":28,"title":29,"summary":30,"content":31,"status":32,"tags":33,"seo_keywords":39,"published_at":45,"updated_at":46,"category":47},16,"geo-technical-pipeline-design","GEO 的技术链路你会怎么设计？","从问题池、证据库、可读页面到答案采样与质量回写，设计一条可维护的 GEO 技术链路，并明确外部 AI 系统的边界。","GEO 的技术链路怎么设计？我会把它当成一套“内容证据的发布与反馈系统”：用真实业务问题组织内容，让公开页面稳定、可核查，再用固定样本观察外部 AI 回答并修正错误。这里不能把外部 AI 的检索和生成过程画成自家可控的服务；不同产品的索引、检索和引用机制并不相同，也不会因为接入某个接口就保证收录或引用。\n\n![GEO 技术链路图：问题池、证据库、发布层、观测层，以及反馈回路](https:\u002F\u002Fnklbhnvyoggxghtaoizu.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fblog-images\u002F2026-10-05-913d451f-e378-4f8b-b825-c93fb4a9a6b3.png)\n\n*图：前三层主要由站点团队负责；外部 AI 如何使用页面只能通过合规采样观察。*\n\n## 先划边界：控制发布，观察回答\n\n自有系统能控制问题选择、事实审核、页面内容、技术可访问性和日志；能观测公开回答中的提及、引用、误述，以及部分引荐与线索。外部平台是否抓取、何时更新索引、如何选取证据、最终怎样措辞，都不能直接保证。因此设计目标不是“让模型按我们的流程回答”，而是让正确资料容易找到、容易理解、容易核对，且发现问题后能快速修订。\n\n如果团队另有自建 RAG 助手，那套系统可以控制切块、向量检索和生成提示。但自建助手的命中率不能直接代表外部 AI 产品中的 GEO 表现，两者应分开监测。\n\n## 第一层：版本化的问题池\n\n先从咨询、销售和站内搜索中整理问题，按用户意图与决策阶段归类。每条问题至少记录：问题 ID、原文、语言与地区、品牌词或非品牌词、用户阶段、来源、负责人和版本。问题池要同时包含“是什么”“怎么选”“风险是什么”等不同类型，避免只测品牌名。\n\n一组问题发布为固定版本后，后续评估尽量沿用同样的题目与运行条件。新增问题另开版本，避免把样本变化误看成内容成效。抽样时不要上传客户原始对话或个人资料；先去标识化，再保留能表达意图的问法。\n\n## 第二层：可追溯的证据库\n\n把需要对外陈述的关键事实独立管理，而不是散落在多篇文章里。每条证据记录主张、来源链接或内部依据、适用范围、生效与复核日期、业务负责人、审核状态，以及引用它的页面。价格、功能边界、交付周期、案例数据尤其需要明确负责人。\n\n以“网站改版会不会影响原有流量”为例，页面可以解释风险条件和检查步骤；如果要写某客户的具体改善数字，就必须找到授权案例和统计口径。证据过期或被撤回时，应能定位所有相关页面，及时修订，不靠编辑逐篇回忆。\n\n## 第三层：把证据发布为可读页面\n\n以 Nuxt + Supabase 站点为例，内容仍由文章系统管理；发布时做一组自动检查：页面返回正常状态码，主要答案在服务端可读的 HTML 中，标题和摘要准确，canonical 指向正式地址，站点地图包含应公开的 URL，robots 与 noindex 没有误拦截，站内链接可用。改版时保留旧 URL 的重定向关系。\n\n结构化数据只能描述页面上真实可见的信息。页面最好直接回答一个主要问题，再写适用条件、步骤、例外和证据来源；相关主题用自然的内链串起来。不要把隐藏文本、关键词堆砌或某个实验性文件当作“被 AI 引用”的保证。\n\n可把发布流程做成四步：编辑提交 → 事实审核 → 页面发布 → 发布后抓取检查。最后一步用实际 URL 验证，而不是只看 CMS 显示“已发布”。对关键页面保存前后版本，出现错误时能回退。\n\n## 第四层：答案采样与质量判定\n\n定时用固定问题集检查目标平台，记录平台名称、可见的模型或版本、时间、语言、问题版本、回答文本、引用 URL、运行失败原因。采样方式应使用平台允许的接口或人工核查，并遵守其使用规则。每个问题重复运行几次，才能看出答案的波动；原始回答与人工判定要留存，方便复核。\n\n自动规则可以初筛“是否提及品牌”“是否引用自有域名”，但“引用是否支持关键说法”“是否把能力或价格说错”需要人工抽查。统计时使用同一套分母和规则；具体公式可参考[《GEO 的核心指标你怎么理解？怎么算？》](\u002Fblog\u002Fgeo-core-metrics-calculation)。引用错误率、失效链接和关键事实误述，应比单纯提及次数更早进入告警列表。\n\n## 最后用反馈推动修订\n\n观测结果要能落到具体动作：没有被提及时，检查问题覆盖、页面入口和可访问性；被引用但说错时，核对证据、过期页面与相互冲突的表述；有访问却没有有效咨询时，再看落地页是否回答了下一步问题。修改后记录页面版本与时间，等待足够的再观察窗口，不假设外部平台会立即更新。\n\n最小实现不需要复杂平台：一份版本化问题表、一份证据清单、现有文章 CMS、发布后页面检查任务、定期答案采样记录和一张人工复核看板就够了。先让“哪条主张来自哪里、发布在哪页、被怎样引用、谁负责修正”能够追踪，再逐步自动化。GEO 技术链路的价值，正是把内容生产变成可审计、可维护的循环。","published",[34,36],{"name":35,"slug":35},"工程",{"name":37,"slug":38},"GEO","geo",[40,41,42,43,44],"GEO技术链路","GEO技术方案","生成式引擎优化架构","AI搜索内容工程","GEO监测","2026-10-05T03:44:34.947+00:00","2026-10-05T03:44:38.369157+00:00",{"id":48,"name":49,"slug":49},3,"技术笔记",{"items":51,"total":219,"page":4,"pageSize":220},[52,69,85,101,119,137,153,159,174,187,202],{"id":53,"slug":54,"title":55,"summary":56,"status":32,"tags":57,"seo_keywords":61,"published_at":67,"category":68},22,"llm-output-instability-hallucination","模型返回不稳定 \u002F 幻觉怎么办？","区分随机波动与事实幻觉，沿着样本评测、证据检索、结构校验、业务规则和人工回退建立可验证的大模型输出链路。",[58,59],{"name":35,"slug":35},{"name":60,"slug":60},"大模型",[62,63,64,65,66],"模型幻觉怎么办","大模型输出不稳定","LLM幻觉","RAG事实校验","模型评测","2026-10-05T04:17:13.315+00:00",{"id":48,"name":49,"slug":49},{"id":70,"slug":71,"title":72,"summary":73,"status":32,"tags":74,"seo_keywords":77,"published_at":83,"category":84},21,"llm-cost-budget-guardrails","大模型调用怎么控制成本？把预算控制放进请求链路","从请求准入、原子预留、真实用量结算到超支降级，设计一套能在并发与重试下运行的大模型调用预算控制链路。",[75,76],{"name":35,"slug":35},{"name":60,"slug":60},[78,79,80,81,82],"大模型调用成本控制","LLM预算管理","Token费用","模型调用限额","API成本监控","2026-10-05T04:14:07.621+00:00",{"id":48,"name":49,"slug":49},{"id":86,"slug":87,"title":88,"summary":89,"status":32,"tags":90,"seo_keywords":93,"published_at":99,"category":100},20,"llm-structured-output-reliability","结构化输出怎么保证模型按格式返回？","模型返回 JSON 不等于系统拿到可用数据。本文区分提示词、JSON 模式与 Schema 约束，说明如何用服务端校验、有限重试和业务规则构建可靠的结构化输出链路。",[91,92],{"name":35,"slug":35},{"name":60,"slug":60},[94,95,96,97,98],"大模型结构化输出","JSON Schema","模型按格式返回","LLM输出校验","约束解码","2026-10-05T04:10:02.025+00:00",{"id":48,"name":49,"slug":49},{"id":102,"slug":103,"title":104,"summary":105,"status":32,"tags":106,"seo_keywords":111,"published_at":117,"category":118},19,"geo-vs-seo-explained","理解 GEO 是什么？和 SEO 有什么区别？","GEO 是什么，和 SEO 有什么区别？从用户提问、内容呈现、优化动作与衡量指标四个角度讲清两者的关系，并给出一套适合建站团队的入门做法。",[107,110],{"name":108,"slug":109},"建站","jianzhan",{"name":37,"slug":38},[112,113,114,115,116],"GEO是什么","GEO和SEO的区别","生成式引擎优化","SEO","AI搜索优化","2026-10-05T04:07:05.673+00:00",{"id":48,"name":49,"slug":49},{"id":120,"slug":121,"title":122,"summary":123,"status":32,"tags":124,"seo_keywords":129,"published_at":135,"category":136},18,"how-nuxt-ssr-works","Nuxt 的 SSR 是怎么做的？从请求、数据获取到 Hydration","沿着一次文章页请求，看 Nuxt 如何用 Nitro 接收请求、在服务端执行 useFetch 并生成 HTML，再由浏览器复用 payload 完成 Hydration。",[125,126],{"name":35,"slug":35},{"name":127,"slug":128},"Nuxt","nuxt",[130,131,132,133,134],"Nuxt SSR","Nuxt服务端渲染","Nuxt useFetch","Hydration","Nitro渲染流程","2026-10-05T03:55:36.035+00:00",{"id":48,"name":49,"slug":49},{"id":138,"slug":139,"title":140,"summary":141,"status":32,"tags":142,"seo_keywords":145,"published_at":151,"category":152},17,"llm-api-cost-control","大模型调用怎么控制成本？一笔账算清缓存、输出与重试","先按输入、输出和额外调用算清单次成本，再用缓存、模型路由、上下文裁剪和预算限额降本，并用可用答案质量守住效果。",[143,144],{"name":35,"slug":35},{"name":60,"slug":60},[146,147,148,149,150],"大模型调用成本","LLM API成本控制","Token费用计算","提示词缓存","模型路由","2026-10-05T03:49:26.666+00:00",{"id":48,"name":49,"slug":49},{"id":27,"slug":28,"title":29,"summary":30,"status":32,"tags":154,"seo_keywords":157,"published_at":45,"category":158},[155,156],{"name":35,"slug":35},{"name":37,"slug":38},[40,41,42,43,44],{"id":48,"name":49,"slug":49},{"id":160,"slug":161,"title":162,"summary":163,"status":32,"tags":164,"seo_keywords":166,"published_at":172,"category":173},15,"geo-core-metrics-calculation","GEO 的核心指标你怎么理解？怎么算？","用同一组问题和明确的分母，算清 GEO 的品牌提及率、引用率、有效引用率、竞争提及份额与有效线索率。",[165],{"name":37,"slug":38},[167,168,169,170,171],"GEO核心指标","GEO指标怎么算","品牌提及率","AI引用率","生成式引擎优化效果","2026-10-05T03:40:09.175+00:00",{"id":48,"name":49,"slug":49},{"id":175,"slug":176,"title":177,"summary":178,"status":32,"tags":179,"seo_keywords":182,"published_at":185,"category":186},14,"geo-business-understanding","GEO 从业务理解开始：把真实问题写成可引用的内容","做 GEO 之前，先弄清业务服务谁、用户在什么场景下提问、哪些证据能支持回答，再用真实问题和业务结果验证内容。",[180,181],{"name":108,"slug":109},{"name":37,"slug":38},[37,114,183,184,116],"业务理解","GEO内容策略","2026-10-05T03:33:52.358+00:00",{"id":48,"name":49,"slug":49},{"id":188,"slug":189,"title":190,"summary":191,"status":32,"tags":192,"seo_keywords":194,"published_at":200,"category":201},13,"website-launch-checklist","网站上线前后检查清单：从可访问到可维护","一份实用的网站上线清单，逐项检查域名、HTTPS、手机体验、页面速度、SEO、表单、备份与上线后的监测。",[193],{"name":35,"slug":35},[195,196,197,198,199],"网站上线检查清单","网站上线","网站性能优化","网站SEO","网站维护","2026-10-05T03:24:44.969+00:00",{"id":48,"name":49,"slug":49},{"id":203,"slug":204,"title":205,"summary":206,"status":32,"tags":207,"seo_keywords":209,"published_at":215,"category":216},12,"website-planning-checklist","建站之前：用一张页面清单理清目标、结构与内容","从网站目标、访问者任务到页面清单和内容准备，用一套可执行的方法完成建站规划，避免边做边改。",[208],{"name":35,"slug":35},[210,211,212,213,214],"建站规划","网站信息架构","网站内容规划","页面清单","个人网站建设","2026-10-05T03:23:27.683+00:00",{"id":217,"name":218,"slug":218},7,"项目实践",11,100,[],1791174017678]