{"version":"https://jsonfeed.org/version/1.1","title":"OnlyFDE","home_page_url":"https://www.onlyfde.cn","feed_url":"https://www.onlyfde.cn/feed.json","description":"制造企业的人在搜索引擎或 AI 里问「工厂怎么用 AI」时，希望出现的那个页面。","language":"zh-CN","authors":[{"name":"OnlyFDE 编辑部","url":"https://www.onlyfde.cn"}],"items":[{"id":"https://www.onlyfde.cn/scenarios/rd-sor-parsing","url":"https://www.onlyfde.cn/scenarios/rd-sor-parsing","title":"把 SOR 解析交给 AI 先做一遍，三百页规格书新人两周接得住","summary":"客户给的 SOR 动辄两三百页：术语密、缩写多，资格认证、禁限用这类门槛条款散落在十几个章节里。新人读不动，老人没空带。本文写给研发型工厂的老板——AI 把规格书拆成结构化需求清单、术语表和门槛标注之后，新人接手能快多少、哪些工程判断仍要老师傅做、一年接几个项目才值得上。","content_text":"SOR 解析，是用 AI 把客户提供的两三百页需求规格书（SOR）拆成带原文页码的结构化需求清单、术语表和资格认证禁限用门槛标注三样交接材料，使没参与前期沟通的新人也能按清单接手项目。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-10-12T09:00:00+08:00","tags":["SOR 解析","需求规格书解析","术语表","门槛条款","研发交接"],"_onlyfde":{"stage":"rd","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/scenarios/nonstandard-ai-dilemma","url":"https://www.onlyfde.cn/scenarios/nonstandard-ai-dilemma","title":"为什么标准化的 AI 产品很好，装到非标工厂里就跑不动","summary":"同一个 AI，写周报又快又好，读一份 SOR 却漏掉强制性条款。本文拆解「非标」到底是什么、标准化 AI 产品缺的两块东西、判断一个方案能不能进工厂的三个问题，以及什么样的厂其实不用折腾这件事。","content_text":"非标制造 AI 的困局，是指在 SOR、图纸、技术协议等核心文档语义密度高、隐性知识多、容错率为零的非标制造企业里，通用大模型和标准化产品读不准文档、没人敢用，AI 只能停留在办公室场景。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-10-12T09:00:00+08:00","tags":["非标制造 AI","工业 AI 落地","通用大模型进工厂","前沿部署工程师模式","工业知识图谱"],"_onlyfde":{"stage":"general","type":"scenario","collection":"scenarios","readingTime":6}},{"id":"https://www.onlyfde.cn/scenarios/material-nl-inventory-query","url":"https://www.onlyfde.cn/scenarios/material-nl-inventory-query","title":"在微信群问一句库存就有答案，自然语言查库存背后要打通什么","summary":"销售在客户现场问一句「某物料还有多少」，过去要等有人查 ERP、截图、发群里，一来一回一两个小时。自然语言查库存让提问直接变成查询，几秒回带数字的答案。本文写给决定要不要做的人：背后要打通哪几笔账、群里谁有资格问出什么、账实不符时答的为什么也是错数。","content_text":"自然语言查库存，是让员工在微信群等日常对话入口用口语提问库存，系统自动查询现有库存、采购在途和工单占用，并按提问人的权限在数秒内返回带具体数字的可用库存答案。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-10-12T09:00:00+08:00","tags":["自然语言查库存","微信群查库存","库存查询机器人","可用库存","物料库存查询"],"_onlyfde":{"stage":"material","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/scenarios/fulfillment-order-breakdown","url":"https://www.onlyfde.cn/scenarios/fulfillment-order-breakdown","title":"非标订单拆解原来靠老师傅，AI 拆任务、节点和责任人靠不靠谱","summary":"非标订单一单一议，拆任务、排节点、定责任人全靠老计划员的经验，他一休假全厂慢半拍。本文写给决定要不要做这件事的人：AI 按历史订单和工艺路线能拆到什么程度、哪里一定会拆错、复核谁来担、没有老师傅的厂能不能起步。","content_text":"非标订单拆解，是用 AI 参照历史订单和工艺路线，把一单一议的非标订单拆成带前后依赖、责任岗位和节点工期的任务初稿，再由计划员复核特例工艺和客户口头约定后用于排产、派工和进度跟踪。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-10-12T09:00:00+08:00","tags":["非标订单拆解","非标订单排产","工艺路线","订单任务分解","生产计划"],"_onlyfde":{"stage":"fulfillment","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/scenarios/document-deep-processing","url":"https://www.onlyfde.cn/scenarios/document-deep-processing","title":"什么是深度文档处理，为什么它是工业 AI 进工厂的第一道门槛","summary":"SOR、图纸、招标文件、技术协议——工业业务跑在这批文档上，AI 要进业务流，先得把它们读准。本文讲清深度文档处理和「AI 读 PDF」的区别、为什么它是各环节 AI 应用的地基，以及怎么用你们的真实文档验证供应商是不是真有这个能力。","content_text":"深度文档处理，是用 AI 把 PDF、扫描件、图纸、表格等工业文档还原成带结构、带原文出处、可逐条回溯核对的条款与参数清单，让报价、工艺、质检等下游业务能直接使用，区别于只识别文字的 OCR。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-10-12T09:00:00+08:00","tags":["深度文档处理","智能文档处理","工业文档解析","文档结构化","SOR 解析"],"_onlyfde":{"stage":"document","type":"scenario","collection":"scenarios","readingTime":6}},{"id":"https://www.onlyfde.cn/scenarios/bidding-document-parsing","url":"https://www.onlyfde.cn/scenarios/bidding-document-parsing","title":"上百页的招标文件解析，AI 几分钟做完，废标风险列得全吗","summary":"招标文件几十上百页，资格条款和废标条款分散在不同章节，漏一条就白做一本标书。本文回答：AI 读标到底读到什么程度、哪些判断它替不了人、什么样的投标团队适合先用，以及怎么用过去的投标记录判断这件事值不值。","content_text":"招标文件解析，是用 AI 把一份几十到几百页的招标文件自动拆成资格、商务、技术三类条款清单，并逐条标注原文页码和废标风险，使投标团队在拿到文件当天就能判断能不能投、谁来投。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-10-12T09:00:00+08:00","tags":["招标文件解析","AI 读标","废标风险","投标决策","资格条款"],"_onlyfde":{"stage":"bidding","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/guides/guide-what-is-fde-forward-deployed-engineer","url":"https://www.onlyfde.cn/guides/guide-what-is-fde-forward-deployed-engineer","title":"什么是 FDE（前沿部署工程师）：制造业为什么需要懂业务又懂 AI 的驻场工程师","summary":"制造业上 AI 最大的困境不是算法不行，而是懂代码的人不懂产线、懂产线的人不会用 AI。FDE（前沿部署工程师）源自军工与前沿数据工程，通过扎进车间现场、与业务人员同坐，用真实业务结果验证 AI 价值。本文给出一张四种交付模式对比表、三条判断规则与避坑指南。","content_text":"FDE（Forward Deployed Engineer，前沿部署工程师，也译作前线部署工程师）指直接派驻到客户业务第一线的高级技术专家。他们兼具系统工程能力与深度的工业现场理解，不按人天售卖工时，也不只写方案，而是深入车间与业务人员共同定位瓶颈，用小步快跑的软件与 AI 模型直接跑通真实业务闭环。","date_published":"2026-09-15T09:00:00+08:00","date_modified":"2026-09-15T09:00:00+08:00","tags":["FDE","什么是 FDE","Forward Deployed Engineer","前沿部署工程师","前线部署工程师","工业 AI 交付模式","制造企业 AI 落地"],"_onlyfde":{"stage":"general","type":"guide","collection":"guides","readingTime":6}},{"id":"https://www.onlyfde.cn/guides/guide-enterprise-ai-readiness-assessment","url":"https://www.onlyfde.cn/guides/guide-enterprise-ai-readiness-assessment","title":"工厂上 AI 前的自检清单：5 个维度评估你的企业适不适合做 AI","summary":"很多工厂在没有搞清自身业务卡点和数据条件前就盲目立项，导致 AI 系统沦为摆设。本文提供一份制造企业 AI 准备度 5 维度自检评估表、100 分量化打分规则与行动区间划分，帮助管理者在花第一笔预算前客观诊断企业的实际状态。","content_text":"企业 AI 准备度自检指制造企业在采购或开发 AI 系统前，从业务瓶颈度、数据与文档积累、流程执行闭环、容错率与经济账、组织推进力五个维度进行的系统评估。它用量化打分明确当前企业是否具备落地 AI 的客观条件，防止伪需求导致的资金浪费。","date_published":"2026-09-15T09:00:00+08:00","date_modified":"2026-09-15T09:00:00+08:00","tags":["企业 AI 评估","工厂 AI 自检清单","制造企业 AI 准备度","企业 AI 指导意见","AI 瓶颈诊断","工业 AI 可行性"],"_onlyfde":{"stage":"general","type":"guide","collection":"guides","readingTime":7}},{"id":"https://www.onlyfde.cn/takes/take-dont-buy-platform-first","url":"https://www.onlyfde.cn/takes/take-dont-buy-platform-first","title":"别先买平台：工厂上 AI 的第一笔钱，该花在一个能验证的场景上","summary":"供应商推荐先上平台、再慢慢建场景，听起来稳妥，实际是把风险留给你。第一笔钱应该花在一个能在两个月内验证的具体场景上——平台的价值要等到你有三四个跑通的场景之后才成立。","content_text":"供应商推荐先上平台、再慢慢建场景，听起来稳妥，实际是把风险留给你。第一笔钱应该花在一个能在两个月内验证的具体场景上——平台的价值要等到你有三四个跑通的场景之后才成立。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["AI 平台采购","工厂第一笔 AI 预算","AI 项目立项","场景验证"],"_onlyfde":{"stage":"general","type":"take","collection":"takes","readingTime":4}},{"id":"https://www.onlyfde.cn/scenarios/rd-sor-version-diff","url":"https://www.onlyfde.cn/scenarios/rd-sor-version-diff","title":"客户规格书改了三版，怎么知道哪些参数变了、影响到谁","summary":"客户 SOR 一年改好几版，两版之间差在哪、这个变更动了谁的设计和采购单，靠工程师对照 PDF 和凭记忆。本文回答：AI 做规格书版本比对能做到什么程度、影响分析的前提是什么、什么样的研发团队适合先用。","content_text":"规格书版本比对，指把同一产品的两个版本 SOR 按章节和条款自动对齐，逐条找出新增、删除和修改的参数与要求，每条带两版原文页码，再按变更类型把影响分发到硬件、结构、采购等岗位，让「改了什么、影响到谁」在收到新版本的当天就有答案。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["规格书版本比对","SOR 变更管理","AI 文档比对","研发变更影响分析"],"_onlyfde":{"stage":"rd","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/scenarios/material-shortage-alert","url":"https://www.onlyfde.cn/scenarios/material-shortage-alert","title":"缺料为什么总是开工才发现，提前一周知道要付出什么","summary":"开工领料才发现缺料，是因为库存、在途、需求三笔账从来不在同一天对。本文回答：缺料预警怎么提前 5–7 天列出缺什么、缺多少、采购赶不赶得上，以及它的前提——每天把三笔账对齐——要付出什么成本。","content_text":"缺料预警，指把工单需求按 BOM 展开、按开工日期倒排，与现有库存和在途采购单每天对齐，提前数天列出未来要开工的工单缺什么料、缺多少、哪天必须要、对应采购单赶不赶得上，让缺料在开工前而不是开工当天暴露。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["缺料预警","物料齐套检查","采购交期跟踪","停线损失"],"_onlyfde":{"stage":"material","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/scenarios/general-first-ai-budget","url":"https://www.onlyfde.cn/scenarios/general-first-ai-budget","title":"工厂 AI 预算第一笔钱该花在招投标研发履约还是物料环节","summary":"老板决定掏出第一笔 AI 预算，招投标、研发、履约、物料四个环节都觉得自己最急。本文回答第一笔钱该按什么顺序花：按数据现成程度和算得清的旧账排序，投给 90 天内能用真实数据验证的单点场景，以及什么时候这笔钱应该先别花。","content_text":"工厂 AI 预算的投法，指企业第一笔 AI 投入不平均分给各环节、也不按供应商的演示决定，而是按「数据现成程度 × 算得清的旧账」给候选场景排序，把第一笔钱投给能在 90 天内用真实数据验证价值的单点场景，验证后再决定下一笔。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["工厂 AI 预算","AI 预算","工厂 AI 落地顺序","AI 试点场景","单点场景验证"],"_onlyfde":{"stage":"general","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/scenarios/fulfillment-bottleneck-view","url":"https://www.onlyfde.cn/scenarios/fulfillment-bottleneck-view","title":"老板不开会，怎么知道全公司订单现在卡在哪","summary":"订单进度靠层层问、周会靠各部门对 Excel，延期常常临近交期才暴露。本文回答：订单卡点看板怎么把每张订单的节点状态摊开、它的上限为什么是现场录入纪律、没有 MES 的厂能不能先用起来。","content_text":"订单卡点看板，指把订单、工单、采购和报工数据汇到一处，把每张订单拆成若干节点，实时显示哪张订单卡在哪个节点、卡了几天、离承诺交期还有几天，超过阈值自动标红上浮，让管理层不开会就能看到全公司的履约卡点。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["订单卡点","订单进度看板","生产瓶颈预警","订单延期预警"],"_onlyfde":{"stage":"fulfillment","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/scenarios/energy-power-ledger","url":"https://www.onlyfde.cn/scenarios/energy-power-ledger","title":"电费账单只有一个总数，用能账本怎么拆到产品批次和班组","summary":"电费每月只有一个总数，哪个产品、哪个班组耗电多没人说得清，报价靠产量粗摊、考核凭印象。本文写给老板：用能账本把电费拆到产品、批次和班组，讲清拆完之后能拿回什么、没有分表计量的厂做到哪一步、什么时候不该做。","content_text":"用能账本是一种工厂用电成本核算方法：把月度电费按设备和生产时段拆分，再归集到产品、批次和班组，让每张订单的用电成本可查、可算、可对账，为报价核算、节能立项和班组考核提供同一套数字依据。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["用能账本","电费分摊","工厂能耗核算","产品用电成本","班组能耗考核"],"_onlyfde":{"stage":"energy","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/scenarios/energy-peak-valley-scheduling","url":"https://www.onlyfde.cn/scenarios/energy-peak-valley-scheduling","title":"峰谷电价差拉到 3 倍以上，工厂的大功率工序怎么排进低谷时段","summary":"分时电价把一天切成尖峰、峰、平、谷四段，政策要求价差不低于 3:1。同样的产量，排在什么时段决定电费差多少。本文写给决定要不要做这件事的人：能拿回多少、什么样的厂适合、什么时候不该做。","content_text":"峰谷电价排产，是在不改变生产总量、工艺参数和交期的前提下，把熔炼、热处理、烘干、空压这类大功率工序的运行时段，从电价高的尖峰段和峰段移到平段和谷段，用同样的用电量换更低的电费。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["峰谷电价排产","分时电价","工厂错峰用电","大功率工序排产","尖峰电价"],"_onlyfde":{"stage":"energy","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/scenarios/energy-idle-standby-waste","url":"https://www.onlyfde.cn/scenarios/energy-idle-standby-waste","title":"下游停了上游还在烧电：高耗能工序的空等每年烧掉多少钱","summary":"熔炼、热处理、烘干、空压这些设备，在下游停机、待料、换型时照样保温和空载运行，产出为零，电表照走。本文写给决定要不要动这件事的人：一年大概烧掉多少、能拿回多少、为什么它通常该排在第一个做。","content_text":"高耗能工序空等，指熔炼、热处理、烘干、空压这类大功率设备，在下游产线停机、待料或换型期间仍维持保温、待机或空载运行所消耗的电——这段时间产出为零，电表照走。它是工厂用电里最容易识别、也最常年被漏掉的一块浪费。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["高耗能设备空转","待机能耗","空压机空载","工序空等","用能账本"],"_onlyfde":{"stage":"energy","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/scenarios/energy-demand-response","url":"https://www.onlyfde.cn/scenarios/energy-demand-response","title":"电网发来需求响应邀约，工厂怎么在半小时内算清该不该应","summary":"需求响应是电网付钱请你少用一会儿电，但邀约通常只给几十分钟决策时间，多数工厂来不及算就放弃了。本文写给决定要不要参与的人：能赚多少、什么样的厂够格、什么时候不该应。","content_text":"电力需求响应，是电网在供需紧张或新能源大发时向用电企业发出邀约，企业按约定在指定时段削减或增加用电负荷，事后按核定的实际调节量获得补偿的机制。对工厂而言，它是用短时让出可调节负荷换来的一笔额外收入。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["电力需求响应","需求响应邀约","可调节负荷","削峰填谷","虚拟电厂"],"_onlyfde":{"stage":"energy","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/scenarios/bidding-response-framework","url":"https://www.onlyfde.cn/scenarios/bidding-response-framework","title":"招标文件读完之后，AI 生成的投标响应框架能不能直接当标书用","summary":"读完招标文件只是第一步，标书还得人来写。本文写给要批这笔投入的老板：AI 生成的投标响应框架能替代标书的哪部分、价格和承诺为什么必须人定、历史标书进知识库能省多少、什么样的团队做了不划算，并用一份假设账演示一年能省出多少人天。","content_text":"投标响应框架，是 AI 读完招标文件后生成的一套标书结构与分工：目录编排、每章要响应的条款、缺失资质清单、各项任务的责任人和截止时间，并从历史标书知识库推送可复用的段落。它替代的是搭骨架和翻旧稿，不替代定价、技术承诺和最终审核。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["投标响应框架","标书框架生成","历史标书复用","投标任务分工","AI 写标书"],"_onlyfde":{"stage":"bidding","type":"scenario","collection":"scenarios","readingTime":5}},{"id":"https://www.onlyfde.cn/guides/guide-geo-industrial-website","url":"https://www.onlyfde.cn/guides/guide-geo-industrial-website","title":"客户都去问 AI 了，工业企业官网怎么被 AI 搜索看到、被整段引用","summary":"采购工程师找供应商，越来越多先问 AI。多数工业企业的官网对 AI 几乎隐形：参数在画册 PDF 里、没有 FAQ、没有结构化数据。本文给一张 GEO 与 SEO 的对比表、一套一周能做完的自检清单，和三个月的起步节奏。","content_text":"官网 GEO（生成式引擎优化），指让工业企业的官网内容能被生成式 AI 准确理解、引用和复述的一系列做法：把产品参数从图片和 PDF 里搬成网页文本、补齐结构化数据、建设 FAQ、给关键事实标注来源，让 AI 在回答采购类问题时提到你、且说得对。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["官网 GEO","生成式引擎优化","AI 搜索优化","工业企业官网","llms.txt"],"_onlyfde":{"stage":"general","type":"guide","collection":"guides","readingTime":6}},{"id":"https://www.onlyfde.cn/guides/guide-deployment-modes","url":"https://www.onlyfde.cn/guides/guide-deployment-modes","title":"公共云、云端私有、本地私有、企业私有：制造企业的 AI 该部署在哪","summary":"AI 部署在哪，决定了数据走不走出厂、要不要买 GPU、出了问题谁来修，而且改起来很贵。本文给一张四种部署方式的对比表和一套判断规则，帮你在签合同之前就把这件事定对。","content_text":"制造企业的 AI 部署方式，指模型和数据运行在什么地方，通常分四档：公共云、云端私有、本地私有、企业私有。选哪一档取决于数据能不能出厂、有没有 IT 人手、以及愿不愿意为算力先花一笔硬件钱。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["AI 私有化部署","云端私有部署","本地私有部署","制造企业 AI 部署","数据不出厂"],"_onlyfde":{"stage":"general","type":"guide","collection":"guides","readingTime":7}},{"id":"https://www.onlyfde.cn/guides/guide-choose-vendor","url":"https://www.onlyfde.cn/guides/guide-choose-vendor","title":"几家工业 AI 供应商比来比去，签合同前必须问清的八个问题","summary":"面对 2–3 家工业 AI 供应商，演示效果都差不多，真正的差别藏在报价口径、行业案例、交付后维护和数据条款里。本文给一张三类供应商的对比表、八个能当面问出口的问题，以及签合同前必须写死的六件事，帮你把风险提前问清。","content_text":"工业 AI 供应商指为制造企业提供 AI 系统或能力的乙方，常见三类：通用大模型厂商、垂直行业软件商、定制开发团队。选型的关键不在演示效果，而在报价口径、真实行业案例、交付后的维护责任和数据归属条款。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["工业 AI 供应商","AI 供应商选型","工业 AI 合同","AI 项目验收"],"_onlyfde":{"stage":"general","type":"guide","collection":"guides","readingTime":7}},{"id":"https://www.onlyfde.cn/guides/guide-ai-project-failure-modes","url":"https://www.onlyfde.cn/guides/guide-ai-project-failure-modes","title":"AI 项目失败最常见的三种死法，签字立项前就能看出来","summary":"车间里的 AI 项目死法惊人地一致：场景是拍脑袋选的、数据是开工后才理的、上线后没人用没人管。本文把三种最常见的死法拆开，给一张含死因、早期信号和补救可能的对比表，再给三条立项签字前的判断规则，帮你在预算花出去之前就看出这个项目会不会死。","content_text":"AI 项目失败，指投入预算和人力之后，系统在一年内停工、废弃或退回人工流程，钱花了但没有形成日常使用。最常见的三种死因是：场景没经过真实数据验证、数据没理顺就开工、上线后没人使用没人维护。三种死因在立项阶段都有可识别的信号。","date_published":"2026-09-06T09:00:00+08:00","date_modified":"2026-09-06T09:00:00+08:00","tags":["AI 项目失败","AI 项目立项","工业 AI 落地","AI 项目复盘"],"_onlyfde":{"stage":"general","type":"guide","collection":"guides","readingTime":7}}]}