舊 V 網歷史統計

這份快照保留搬站前可公開讀取的歷史數字。來源是舊 V 網的站內 Upstash 計數器,與 Vercel Web Analytics 平台報表分開計算。

擷取時間: 至 (台北 UTC+8)。擷取當下原站仍在服務;這份快照不會隨新流量更新。

累計 10,104 次瀏覽

這是舊算法記錄的瀏覽次數,不是獨立人數。原前端對同瀏覽器、同頁一小時去重,後端排除已識別 bot,另有同 IP 每分鐘寫入上限;不是所有爬蟲或所有造訪都能被辨識。每日、每月依台北時區分桶。

不要重複加總:C 網正式接上同一份 Upstash 後,原總數已包含這些歷史;快照只供追溯,不另外加進新總數。Cloudflare 平台流量也不能直接和這些數字合併。

歷月瀏覽次數
月份次數狀態
2026-073,070舊 API 歷月記錄
2026-082,840舊 API 歷月記錄
2026-093,847舊 API 歷月記錄
2026-10347擷取當月,未結算
擷取當月的每日記錄
台北日期次數
2026-10-01141
2026-10-0264
2026-10-0378
2026-10-0464
查看 231 個公開頁路徑的累計數字

逐頁依序讀取,與全站總數不是同一瞬間的快照。清單來自本次網站公開 HTML,未涵蓋所有歷史別名或已刪頁的資料鍵;請勿用此表加總代表全站。0 只代表舊 API 回傳 0,該 API 無法區分缺值與真零。

公開頁路徑累計
路徑累計次數
/1,876
/agent.html3
/ai-office/131
/ai-office/assistants.html0
/ai-office/learn-handbook.html0
/ai-office/learn-slides.html7
/ai-office/learn.html17
/ai-office/lyca/22
/ai-office/team.html27
/ai-power-tower/24
/ai-trends.html9
/ai-trends/apple-wwdc26-siri-lobster-ai/77
/ai-trends/demis-hassabis-agi-science-ai/8
/ai-trends/program-vs-ai-skill-library/6
/ai-trends/spacex-ipo-musk-trillionaire-knowledge-work/5
/articles.html455
/articles/a2a-agent-protocol/5
/articles/after-ai-says-remembered/11
/articles/agent-native-tools-software-interface/5
/articles/agent-web-turning-point/12
/articles/agent-workflow-builds-automation/9
/articles/agi-work-and-discovery/5
/articles/ai-adoption-raise-not-install/0
/articles/ai-adoption-start-with-one-workflow/0
/articles/ai-batch-work-without-seeing-data/0
/articles/ai-capability-tiers/14
/articles/ai-changed-behavior-into-workflow/8
/articles/ai-course-map-from-entry-to-workflow/19
/articles/ai-cp-value-calculus/14
/articles/ai-custom-work-compound/0
/articles/ai-customer-service-api-cost/0
/articles/ai-data-organization-usable-system/20
/articles/ai-data-red-lines/0
/articles/ai-delegators-optimism/2
/articles/ai-employee-four-levels/7
/articles/ai-employees-manage-agents/0
/articles/ai-era-websites-for-agents/0
/articles/ai-handoff-instructions/3
/articles/ai-image-api-cost/0
/articles/ai-king-off-battlefield/4
/articles/ai-learning-map-reduce-anxiety/15
/articles/ai-loop-safety-recovery/6
/articles/ai-market-microcosm/18
/articles/ai-mvp-validation-before-product/4
/articles/ai-native-not-transformation/4
/articles/ai-rule-file-slimming/0
/articles/ai-said-it-watched-the-video/8
/articles/ai-schedule-wake-interval/0
/articles/ai-slides-half-auto-google-slides/0
/articles/ai-that-knows-you/5
/articles/ai-tools-professional-judgment/5
/articles/ai-understands-ei-ha/0
/articles/ai-usage-audit/0
/articles/ai-user-testing/0
/articles/ai-vs-script-four-situations/0
/articles/answer-in-person-or-ai/1
/articles/answer-to-action-enterprise-ai-agent/6
/articles/audrey-tang-reverse-alignment/0
/articles/before-installing-others-skill/2
/articles/book-to-ai-consultant/5
/articles/books-videos-pdf-ai-advisor/7
/articles/build-your-own-dictionary/2
/articles/cathay-tech-conference-2026/0
/articles/caught-ai-slacking-into-rules/4
/articles/character-costume-sheet-three-views/27
/articles/chatgpt-work-codex-choice/22
/articles/chatgpt-work-skills-web-version/10
/articles/claude-meta-ads-connector/0
/articles/claude-skills-knowledge-assets/64
/articles/cli-api-mcp-computer-use/15
/articles/cloudflare-service-map/0
/articles/codex-log-health-check/125
/articles/codex-only-auto-worklog/32
/articles/cognitive-debt/3
/articles/company-shape-is-the-moat/6
/articles/copied-mechanism-is-no-mechanism/2
/articles/cross-ai-review-both-wrong/12
/articles/decision-ladder-non-programmer/78
/articles/delete-before-automate/0
/articles/design-for-ai-errors/0
/articles/diary-driven-agent-3x4/49
/articles/docs-as-system-design-agent/50
/articles/domain-and-account-ownership/0
/articles/dont-learn-ai-tools/2
/articles/dual-centaur-meeting/33
/articles/dual-track-planning-loop/79
/articles/eight-ai-system-concepts-2026/24
/articles/elon-musk-live-skill/16
/articles/first-principles-money-soup/0
/articles/five-loops-content-line/0
/articles/fixed-test-set-for-ai/0
/articles/four-lens-rapid-review/9
/articles/free-deploy-three-boundaries/27
/articles/github-vercel-cloudflare-compare/31
/articles/give-ai-choices-not-descriptions/41
/articles/grok-bot-galaxy-day1-founders/0
/articles/grok-bot-galaxy-day2-workshops/0
/articles/grok-bot-galaxy-day3-shipping/0
/articles/harness-mindset-for-bosses/16
/articles/harness-to-loop/0
/articles/hosting-platform-decision-tree/0
/articles/how-ai-connects-software/18
/articles/how-to-train-your-ai-employee/44
/articles/how-to-verify-ai-rule-changes/0
/articles/inspiration-production-system/26
/articles/intangible-assets-grow-by-sharing/4
/articles/intent-first-prompting/22
/articles/knowledge-as-employee/4
/articles/knowledge-base-three-vault-split/37
/articles/knowledge-os-master-map/7
/articles/laptop-desktop-webpage-sync-icloud-git/27
/articles/let-ai-do-the-setup/4
/articles/let-ai-review-itself/0
/articles/liberal-arts-agent-framework/16
/articles/line-backup-who-can-see-it/0
/articles/line-group-ai-workflow/109
/articles/llm-rag-agent-mcp/7
/articles/long-answer-three-layers/18
/articles/long-document-review-layers/1
/articles/long-task-completion-rate/9
/articles/loop-engineering-guardrails/14
/articles/loop-four-entries/0
/articles/loop-round-limit/0
/articles/mac-for-agent-beginners/71
/articles/make-your-ai-secretary/0
/articles/manage-ai-with-management-knowledge/5
/articles/map-is-not-the-territory/8
/articles/market-to-ai/14
/articles/meeting-into-verifiable-loop/6
/articles/meeting-mic-rescue-on-mac/0
/articles/meeting-record-agent-workflow/104
/articles/messaging-apps-ai-friendliness/10
/articles/meta-prompt-thirty-versions/2
/articles/mika-to-laika-product-character-design/7
/articles/mobile-to-desktop-publish-loop/4
/articles/my-three-loops/31
/articles/notebooklm-knowledge-analysis-assistant/5
/articles/notebooklm-real-photo-slides/25
/articles/old-prompts-intent-first-loop-engineering/24
/articles/one-on-one-questions/6
/articles/one-sentence-ai-website-workflow/0
/articles/one-sentence-video-to-deep-article/0
/articles/openrouter-deepseek-data-routing/5
/articles/opus-luna-model-division/0
/articles/own-ai-team-at-work/12
/articles/parallel-site-editing/15
/articles/parenting-story-ai-family-dialogue/4
/articles/personal-poster-and-wallpaper/2
/articles/personal-studio-vs-solo-company/0
/articles/pii-shield/0
/articles/post-class-organizing-loop/97
/articles/prompt-to-loop-map/45
/articles/publish-gate/6
/articles/questionnaire-to-slides-agent-workflow/14
/articles/rag-three-retrieval-modes/3
/articles/recovery-over-perfection/2
/articles/reevaluate-models-same-test/0
/articles/relationship-boundary-ai-practice/5
/articles/rule-file-rebound/3
/articles/same-question-different-answers/17
/articles/self-hosted-site-analytics/0
/articles/semantic-rules-before-prompt-templates/13
/articles/session-messaging-reminder-layer/6
/articles/slides-by-audience-sop/0
/articles/start-from-one-persons-problem/0
/articles/start-with-a-wrong-draft/1
/articles/strong-ai-models-knowledge-workflow-road/24
/articles/tag-wiki-method/9
/articles/talent-vs-expertise/1
/articles/teach-ai-not-learn-ai/10
/articles/teacher-prep-knowledge-workflow/17
/articles/ten-year-anchor/1
/articles/three-levels-of-cross-review/154
/articles/tidy-mess-before-consulting/12
/articles/timing-and-forecast/1
/articles/train-your-ai-agent-editor/20
/articles/use-more-ai-not-enough/6
/articles/vibe-coding-ten-half-products/6
/articles/web-chat-ai-vs-desktop-agent/30
/articles/webmcp-day-one/3
/articles/webnode-to-cloudflare-pages/9
/articles/website-as-workstation/0
/articles/what-are-subagents/0
/articles/what-is-graph-engineering/49
/articles/what-is-loop-engineering/262
/articles/where-is-my-server/0
/articles/who-can-see-your-site/7
/articles/why-split-data-into-cards/4
/articles/yongli-ai-workshop-tools-to-workflow/4
/articles/youtube-to-opinion-report-workflow/10
/cases.html115
/cases/digital-chamber-v3.html7
/cases/event-matchmaking.html1
/cases/meeting-insight.html9
/cases/sme-agent-briefing.html0
/cases/social-tools-comparison.html6
/class-countdown.html0
/courses.html710
/courses/2026-05-23-agent-workflow/44
/courses/2026-05-30-yongli-ai-workshop/10
/courses/2026-06-03-1719-ai-employee-course.html31
/courses/2026-06-14-youtube-report/2
/courses/2026-06-27-parenting-storytelling-ai/23
/courses/2026-07-05-slide-to-skill/6
/courses/2026-07-08-ai-agent-editor/14
/courses/2026-07-08-ai-agent-editor/map.html2
/courses/2026-07-12-chatgpt-work/16
/courses/2026-08-02-dual-model-review-loop/1
/courses/2026-08-04-digital-creation-1/164
/courses/2026-08-15-ai-that-knows-you/117
/courses/2026-08-danhai-public-art/34
/courses/2026-08-danhai-public-art/scenes/3
/courses/2026-08-danhai-public-art/showcase/25
/courses/2026-09-12-danhai-public-art/99
/courses/2026-09-12-danhai-public-art/scenes/46
/courses/2026-09-26-ai-site-and-deploy/0
/courses/2026-09-ai-series/1,499
/courses/2026-09-ai-series/lesson-0/28
/courses/ai-employee-deck/6
/courses/mobile-ai-creation-workshop/2
/courses/senior-ai-story-workshop/12
/courses/senior-ai-story-workshop/deck.html0
/invited-talks.html12
/knowledge-architecture.html466
/learn.html129
/offers.html169
/offers/companion/15
/resources.html31
/search.html11
/skills.html465
/stats.html7

下載這份歷史快照 JSON

尚未接入的歷史資料

聊天、搜尋原文、獨立訪客、來源網站、逐頁每日明細仍需受保護的資料庫或管理權限驗收,這裡沒有顯示這些私密資料,也沒有將缺資料填成 0。原有聊天與搜尋有月筆數上限及保存期限;本次公開快照不延長原保存期限,已過期或已截斷資料也不能由快照恢復。

Vercel Web Analytics 已另行備存 API 可取得的歷史資料,保留在管理用備存檔,尚未接入這個頁面的受保護儀表板。細分資料受方案與查詢範圍限制,不代表完整匯出;Speed Insights 尚未匯出。Vercel 官方 API 說明列出平台統計與查詢限制。

下方密碼入口進入受保護的即時儀表板。2026-10-09 起官網改由 Cloudflare 提供,即時儀表板讀的是同一份 Upstash 原有資料,搬家前後的計數接在一起;上方快照是搬家前唯讀保存的舊站數字。

站長儀表板

這頁要密碼。

到

 

名片分享紀錄(誰從後台分享、按了哪顆按鈕)在另一個後台,密碼不同:開啟名片追蹤後台。名片頁 /card/ 的瀏覽次數會出現在下方熱門頁。

流量趨勢

怎麼讀:看形狀不看單日數字。上面是每天有多少人看網頁,下面是每天有多少人跟咪卡講話。兩張圖分開畫(不共用一個 Y 軸),因為量級差十倍以上,疊在一起看不出小的那條在動。發文、開課、講座之後這兩條會一起翹起來。

哪些頁面被看

怎麼讀:這是從開站到現在的累計,不受上面的期間選擇影響。條越長看的人越多。點任何一列,下面的「咪卡對話明細」會只留這一頁的提問,看得到讀者在這頁卡在哪。

看完整清單(含冷門頁)

這段期間誰在動

怎麼讀:這張只算你選的期間,跟上面的累計榜互補。累計榜是老本,這張是現在的動能。長條是期間內的瀏覽數,灰字是開站到現在的累計。每頁每日的記錄從 2026-09-01 才開始,在那之前沒有這個維度。

咪卡客服:大家都問什麼

怎麼讀:格子越深表示在那一頁問那類問題的人越多。點任何一個格子,下面的明細會只留那一格的對話。問題分類是關鍵詞規則自動分的,不是 AI 判讀,分不出來的一律歸「其他」,原句在最下面的明細都查得到。

要注意的事

怎麼讀:這區是機器用規則掃出來的異常訊號,不是結論。每條都附可以直接查的原句,看過才算數。

咪卡對話明細

怎麼讀:一段對話一張卡,藍色是訪客說的、灰色是咪卡回的,由上往下就是當時的先後順序。上面的圖點下去會只留有命中的那幾段,命中的那句會亮起來。訪客原話一字不改,只看得到瀏覽器識別碼,看不到是誰。

目前:全部

站內搜尋:找不到的東西

怎麼讀:零命中的搜尋詞就是內容缺口或別名缺口。這一項的原始資料是以月為單位存的,期間跨月時是整月相加,切不到單日。詞在站上有對應內容卻搜不到,補 search-aliases.js 的別名;真的沒有內容,那就是下一篇要寫的題目。

人從哪裡來

怎麼讀:只記來源網站的主機名,不記完整網址(完整網址常夾帶查詢字串與個資)。站內互點不算來源。上面那組是從 AI 平台點進來的,這一項直接對應 4O 裡的 GEO 成效:有沒有人因為 AI 提到你而找過來。2026-09-01 才開始記,在那之前是空的。

統計機制自我檢查

怎麼讀:這區在回答「這些數字本身可不可信」。埋點漏一頁,那一頁的流量就是零,看起來像沒人看,其實是沒在數。

記錄方式:自建 Vercel Function + Upstash Redis,不經過 Google。 瀏覽數同一瀏覽器同一頁一小時只計一次;機器人 User-Agent 不計。 咪卡對話只記匿名瀏覽器識別碼,不做會員、不追個人軌跡。