隱私與服務界線Privacy & service boundaries

為健康但不快樂的人，留一個起點。A beginning for people who feel well, yet unhappy.

德派心智教練與慢聊好所在關注許多仍能維持日常生活，卻被情緒、壓力或身體緊繃所困擾的人。我們透過非醫療的身心學習、實體課程、日常分享與 AI 自我反思，陪你覺察感受、表達壓力，探索適合自己的調整與抒發方式。De-Pai mind coaching and The Slow Talk focus on people who manage everyday life but feel weighed down by emotions, stress, or physical tension. Non-medical learning, in-person courses, everyday reflections, and AI self-reflection offer ways to notice feelings, express stress, and explore personal adjustments.

「健康但不快樂」是對服務情境的描述，不是診斷，也不是一個人口比例或成效承諾。AI 試用以成年人為對象，另有使用前適用條件；請閱讀確認後再開始。“Well, yet unhappy” describes an everyday situation, not a diagnosis, a population statistic, or a promise of results. The AI beta is for adults and has additional eligibility conditions to read before beginning.

有健康疑慮時，醫療照顧優先。When health is a concern, medical care comes first.

如果你正在接受身心科、精神科或其他醫療照顧，已有相關診斷，或懷疑自己的身心狀況需要評估，請優先尋求合格醫師或心理專業人員的協助，接受適當且必要的評估與照顧。課程、貼文與 AI 回覆不能判斷你的病情，也不能取代醫療、心理諮商或治療。If you are receiving psychiatric or other medical care, have a relevant diagnosis, or suspect that your physical or mental health needs assessment, seek qualified medical or psychological support first. Courses, posts, and AI replies cannot assess your condition or replace medical care, counselling, or treatment.

請不要因為參與課程或閱讀 AI 回覆而自行停藥、改變治療計畫或延後就醫。若想調整既有照顧，請先與你的醫師或合格專業人員討論，以免影響病情。尊重醫療界線，是 mindful tech 文化的一部分。Do not stop medication, change a treatment plan, or postpone care because of a course or an AI reply. Discuss any change with your doctor or qualified professional first. Respect for medical boundaries is part of our mindful-tech culture.

如果你有立即危險或傷害自己、他人的念頭，請聯絡身邊可信任的人與當地緊急服務。在台灣可撥 1925 安心專線；立即危險請撥 119／110。If you are in immediate danger or thinking of harming yourself or someone else, contact a trusted person and local emergency services. In Taiwan, call 1925 for crisis support; call 119 / 110 in immediate danger.

對話資料如何流動？How does conversation data flow?

對話經 HTTPS 加密傳輸，透過 Cloudflare 與 Anthropic 的 Claude API 產生回覆。本站目前不會將對話內容建立成永久對話資料庫。每次重新開啟對話頁面或開始新對話時，不會載入上一次的對話紀錄，也不會保留跨次對話記憶。本站也不以對話建立本站模型的訓練資料集；這不是加密儲存的個人日記服務。第三方的資料處理、訓練例外與留存規則，請閱讀完整隱私說明。Conversations are transmitted over HTTPS through Cloudflare and Anthropic’s Claude API to generate replies. This site currently does not store conversation content in a permanent conversation database. Each time you open the conversation page anew or start a new conversation, previous chat records are not loaded and no memory carries over from earlier conversations. This site also does not build a training dataset for its own models from your chats; it is not an encrypted journal-storage service. Read the full privacy information for third-party processing, training exceptions, and retention.

當次對話暫存在目前頁面的記憶體中，僅用來維持這一段對話的脈絡；目前沒有個人日記庫或歷史分析功能。重新載入或離開頁面，不等同於刪除第三方平台可能依法或依服務條款留存的資料。The current conversation is temporarily held in the current page’s memory to maintain context within that conversation. There is no saved journal or analysis across separate chats. Reloading or leaving the page does not delete data that third-party platforms may retain under their terms or legal obligations.

網站會在瀏覽器儲存你的閱讀語言偏好，不會用這項設定保存聊天內容。Cloudflare、外部網站與服務可能依各自條款處理必要的連線資訊；我們不承諾零留存或絕對保密。The site stores your reading-language preference in your browser; that setting does not store chat content. Cloudflare and external services may process necessary connection information under their terms. We do not promise zero retention or absolute confidentiality.

模型訓練與第三方留存。Model training and third-party retention.

本站不以對話建立本站模型的訓練資料集。依 Anthropic 的商用資料政策，Claude API 的輸入與輸出預設不供模型訓練；若主動分享回饋、允許資料使用或適用其他安排，則有例外。實際處理以服務條款與帳戶設定為準。This site does not build a training dataset for its own models from your conversations. Anthropic’s commercial policy states that API inputs and outputs are not used for training by default, with exceptions such as explicitly shared feedback, permission to use data, or other applicable arrangements. Actual processing depends on the terms and account settings.

Anthropic 的標準 API 政策通常在 30 天內刪除輸入與輸出，但安全、法律要求、特定功能或契約安排可能有不同留存方式；本站未宣稱已取得零資料留存安排。Anthropic’s standard API policy normally deletes inputs and outputs within 30 days, with exceptions for safety, legal requirements, certain features, or contractual arrangements. This site does not claim a zero-data-retention agreement.

少提供一點，也可以好好聊。You can share less and still reflect.

請避免輸入身分證字號、住址、帳戶資料、可識別他人的資訊或不必要的病歷。你可以用較概括的方式描述感受，隨時暫停對話；有資料處理疑問時，請聯絡 [email protected]，不必在信件中附上敏感的完整對話。Avoid sharing ID numbers, addresses, account details, identifiable information about others, or unnecessary medical records. Describe feelings in general terms and pause whenever you wish. For data questions, contact [email protected] without attaching a sensitive full conversation.

更新日期：2026 年 10 月 10 日Updated: 10 October 2026