Merge pull request #526 from MBombeck/release/v1.30.2

release/v1.30.2
2026-11-08 12:32:43 +00:00 · 2026-07-18 05:26:28 +02:00
commit 3bd550cb21
@@ -2,6 +2,17 @@

 ## [Unreleased]

+## [1.30.2] — 2026-07-18
+
+More audit-backlog fixes.
+
+- **Coach history is fully reachable.** The conversation list pages through your whole history and search runs on the server (by title), instead of only the first ~20 loaded chats.
+- **Navigation and flows.** A document links to the lab values it produced; the two "Recovery" pages are named distinctly and cross-linked; mood, mental wellbeing, and mood insights link to each other; the nutrients card moved out of "Source priority"; the onboarding checklist and tour no longer point at dead ends; asking the coach from a metric or a plan carries that context in.
+- **Correctness.** Correlations, the intraday chart, achievements, and the nutrients overview now bucket days by your own timezone and lead with the most recent data; a multi-source day is no longer counted twice.
+- **Performance.** The workouts list caches its de-duplication per filter, so paging through a long history is quick.
+
+No migration. No breaking changes.
+
 ## [1.30.1] — 2026-07-18

 Polish and small fixes.
@@ -1,7 +1,7 @@
 openapi: 3.1.0
 info:
  title: HealthLog API
-  version: 1.30.1
+  version: 1.30.2
  description: >-
    Self-hosted personal-health-tracking PWA — public API surface for the iOS native client and external ingest.

@@ -5068,8 +5068,10 @@ paths:
      description: v1.18.0 — cursor-paginated list of the caller's Coach conversations for the history rail, most-recent
        activity first. Metadata only (id, title, timestamps, message count); message bodies are not decrypted here.
        `limit` defaults to 20, capped at 50; pass the returned `nextCursor` back as `cursor` for the next page (null at
-        the end). Coach-gated (`requireAssistantSurface("coach")`); a disabled surface 403s. Auth via cookie or Bearer;
-        the owner is narrowed from the session.
+        the end). v1.30.2 — optional `q` narrows the page to conversations whose TITLE contains the text
+        (case-insensitive substring, capped at 200 chars); message bodies are encrypted at rest and are not searched.
+        Coach-gated (`requireAssistantSurface("coach")`); a disabled surface 403s. Auth via cookie or Bearer; the owner
+        is narrowed from the session.
      parameters:
        - name: cursor
          in: query
@@ -5086,6 +5088,14 @@ paths:
            maximum: 50
            default: 20
          description: Page size. Defaults to 20, capped at 50.
+        - name: q
+          in: query
+          required: false
+          schema:
+            type: string
+            maxLength: 200
+          description: v1.30.2 — case-insensitive substring filter over the conversation TITLE only (message bodies are encrypted
+            and not searched). Omit for the unfiltered list.
      responses:
        "200":
          description: A page of the caller's conversations.
@@ -228,7 +228,7 @@
    "logout": "Abmelden",
    "login": "Anmelden",
    "notifications": "Benachrichtigungen",
-    "theme": "Darstellung",
+    "theme": "Design",
    "themeSystem": "System",
    "themeDark": "Dunkel",
    "themeLight": "Hell",
@@ -599,12 +599,13 @@
    },
    "watchdog": {
      "title": "Automatische Überwachung",
-      "description": "HealthLog beobachtet außerdem still im Hintergrund — Benachrichtigungen zu unregelmäßigem Rhythmus und Atemstörungen von einer verbundenen Uhr sowie Warnungen bei Abweichungen von deiner Baseline — und zeigt eine Karte nur, wenn etwas auffällt. Keine Karte ist ein gutes Zeichen, kein fehlendes Feature."
+      "description": "HealthLog beobachtet außerdem still im Hintergrund — Benachrichtigungen zu unregelmäßigem Rhythmus und Atemstörungen von einer verbundenen Uhr, Warnungen bei Abweichungen von deiner Baseline sowie auffällige Veränderungen zwischen deinen letzten beiden Laborwerten — und zeigt eine Karte nur, wenn etwas auffällt. Keine Karte ist ein gutes Zeichen, kein fehlendes Feature."
    }
  },
  "mood": {
    "title": "Stimmung",
    "subtitle": "Stimmungsverlauf und Einträge",
+    "mentalWellbeingLink": "Check-in zum seelischen Wohlbefinden machen",
    "addEntry": "Hinzufügen",
    "editEntry": "Eintrag bearbeiten",
    "moodLevel": "Stimmung",
@@ -2263,6 +2264,8 @@
      "historyEmpty": "Deine Unterhaltungen erscheinen hier. Stell eine Frage, um zu starten.",
      "historyDeleteAria": "Unterhaltung löschen",
      "historyDeleted": "Unterhaltung gelöscht",
+      "historySearchEmpty": "Keine Unterhaltungen entsprechen deiner Suche.",
+      "historyLoadingMore": "Mehr wird geladen…",
      "sourcesTitle": "Worauf ich zugreife",
      "sourcesFresh": "Aktuell",
      "sourcesStale": "Veraltet",
@@ -2894,7 +2897,7 @@
      "sleep": {
        "title": "Noch keine Schlafwerte",
        "description": "Sobald Schlafdaten vorliegen (manuell oder via Withings / Apple Health), erscheint hier der nächtliche Stadien-Verlauf.",
-        "cta": "Withings verbinden"
+        "cta": "Quelle verbinden"
      }
    },
    "navWorkouts": "Workouts",
@@ -3262,6 +3265,8 @@
    },
    "navWalkingSpeed": "Gehgeschwindigkeit",
    "mood": {
+      "logLink": "Stimmung eintragen",
+      "mentalWellbeingLink": "Check-in zum seelischen Wohlbefinden machen",
      "heatmapTitle": "Stimmungskalender",
      "heatmapNoEntry": "Kein Eintrag",
      "legendGreat": "Super",
@@ -3550,7 +3555,8 @@
        "sectionSubtitle": "Zusammengesetzte Werte für Erholung, Stress und Belastung aus deinen verbundenen Geräten",
        "loadError": "Deine Wellness-Werte konnten gerade nicht geladen werden.",
        "loadingLabel": "Deine Gesundheitswerte werden geladen",
-        "recovery": "Erholung",
+        "recovery": "Erholungs-Score",
+        "recoveryCrossLink": "Erholungs- und Belastungsdaten deines Wearables ansehen",
        "stress": "Stress",
        "strain": "Belastung"
      },
@@ -3721,6 +3727,7 @@
    "recovery": {
      "title": "Erholung",
      "description": "Wie gut sich dein Körper erholt hat – die Erholungs-, Belastungs- und Load-Signale, die dein Wearable über Nacht aufzeichnet.",
+      "scoreCrossLink": "Den zusammengesetzten Erholungs-Score ansehen",
      "emptyTitle": "Noch keine Erholungsdaten",
      "empty": "Noch keine Erholungsdaten. Verbinde ein WHOOP-, Polar- oder Oura-Konto, dann füllen sich diese Werte, sobald dein Gerät synchronisiert.",
      "block": {
@@ -4009,7 +4016,7 @@
        },
        "wrapUp": {
          "title": "Das war die Tour",
-          "body": "Du kannst jedes Modul jederzeit einzeln erneut ansehen („Diese Tour zeigen“ auf jeder Seite) oder die komplette Tour unter Einstellungen → Erweitert neu starten."
+          "body": "Diese Führung läuft einmalig – sie öffnet sich nicht von selbst erneut. Alles Gezeigte bleibt erhalten, und du kannst Funktionen jederzeit unter Einstellungen → Module ein- oder ausschalten."
        }
      }
    },
@@ -4084,7 +4091,8 @@
      "learning": "Manche Einblicke – Glukose, Schlafrhythmus, Blutdrucktrends, Zyklus – schärfen sich in den ersten ein bis zwei Wochen, sobald sich deine Daten ansammeln. Leere Kacheln heißen jetzt nur, dass wir deine Basislinie noch lernen.",
      "connectCta": "Gerät verbinden",
      "logCta": "Ersten Wert eintragen",
-      "importCta": "Bestehende Daten importieren"
+      "importCta": "Bestehende Daten importieren",
+      "modulesCta": "Module auswählen"
    },
    "welcomeBack": {
      "title": "Willkommen zurück",
@@ -4549,7 +4557,7 @@
        "description": "Strukturierte Datensätze zu Allergien und Familienanamnese."
      },
      "environment": {
-        "title": "Umwelt",
+        "title": "Standort & Wetter",
        "subtitle": "Wetter, Tageslicht und Standort-Kontext.",
        "disabledHint": "Aktiviere das Umwelt-Modul, um einen Heimatort festzulegen und tägliches Wetter und Tageslicht abzurufen.",
        "searchPlaceholder": "Stadt suchen",
@@ -7959,6 +7967,7 @@
  "mentalHealth": {
    "pageTitle": "Seelisches Wohlbefinden",
    "pageDescription": "Vier validierte, freiwillige Selbsttests — PHQ-9 (Stimmung), GAD-7 (Sorgen), WHO-5 (Wohlbefinden) und SCI (Schlaf) — ergänzend zu deinem Stimmungstagebuch. Aussagekräftig ist der Verlauf über die Zeit, nicht ein einzelner Wert.",
+    "moodLink": "Deinen Stimmungsverlauf ansehen",
    "attributionLabel": "Quelle",
    "attributionShow": "Quelle und Lizenz für {instrument}",
    "choosePrompt": "Check-in auswählen",
@@ -8404,7 +8413,8 @@
      "noConditions": "Mit keiner Erkrankung verknüpft",
      "previewUnavailable": "Keine Inline-Vorschau für diesen Dateityp",
      "uploadedMeta": "Hinzugefügt am {date} · {size}",
-      "share": "Teilen"
+      "share": "Teilen",
+      "labValuesLink": "{count} Laborwerte ansehen"
    },
    "dropOverlay": {
      "title": "Zum Hochladen ablegen",
@@ -8666,7 +8676,9 @@
      "coachCheckin": {
        "title": "Zeit, einen Plan anzuschauen",
        "body": "Es ist etwa eine Woche her, seit du diesen Plan gesetzt hast — behalte ihn, passe ihn an oder lass ihn los. Ganz ohne Druck.",
-        "bodyPlan": "Vor etwa einer Woche hast du dir vorgenommen: „{plan}“. Behalte es, passe es an oder lass es los — ganz ohne Druck."
+        "bodyPlan": "Vor etwa einer Woche hast du dir vorgenommen: „{plan}“. Behalte es, passe es an oder lass es los — ganz ohne Druck.",
+        "adjustPrompt": "Ich möchte meinen Check-in-Plan anpassen.",
+        "adjustPromptPlan": "Ich möchte diesen Plan anpassen: „{plan}“"
      },
      "tensionWindow": {
        "title": "Vielleicht eine angespannte Phase",
@@ -228,7 +228,7 @@
    "logout": "Sign out",
    "login": "Sign in",
    "notifications": "Notifications",
-    "theme": "Appearance",
+    "theme": "Theme",
    "themeSystem": "System",
    "themeDark": "Dark",
    "themeLight": "Light",
@@ -599,12 +599,13 @@
    },
    "watchdog": {
      "title": "Automatic screening",
-      "description": "HealthLog also watches quietly in the background — irregular-rhythm and breathing-disturbance notifications from a connected watch, plus baseline-drift alerts — and surfaces a card only when something stands out. No card showing is a good sign, not a missing feature."
+      "description": "HealthLog also watches quietly in the background — irregular-rhythm and breathing-disturbance notifications from a connected watch, baseline-drift health-status alerts, and notable shifts between your two most recent lab panels — and surfaces a card only when something stands out. No card showing is a good sign, not a missing feature."
    }
  },
  "mood": {
    "title": "Mood",
    "subtitle": "Mood tracker and entries",
+    "mentalWellbeingLink": "Take a mental-wellbeing check-in",
    "addEntry": "Add",
    "editEntry": "Edit entry",
    "moodLevel": "Mood",
@@ -2263,6 +2264,8 @@
      "historyEmpty": "Your conversations will appear here. Ask anything to start.",
      "historyDeleteAria": "Delete conversation",
      "historyDeleted": "Conversation deleted",
+      "historySearchEmpty": "No conversations match your search.",
+      "historyLoadingMore": "Loading more…",
      "sourcesTitle": "What I can see",
      "sourcesFresh": "Fresh",
      "sourcesStale": "Stale",
@@ -2894,7 +2897,7 @@
      "sleep": {
        "title": "No sleep entries yet",
        "description": "Once sleep data lands (manual or via Withings / Apple Health), the per-night stage breakdown will appear here.",
-        "cta": "Connect Withings"
+        "cta": "Connect a source"
      }
    },
    "navWorkouts": "Workouts",
@@ -3262,6 +3265,8 @@
    },
    "navWalkingSpeed": "Walking Speed",
    "mood": {
+      "logLink": "Log a mood entry",
+      "mentalWellbeingLink": "Take a mental-wellbeing check-in",
      "heatmapTitle": "Mood calendar",
      "heatmapNoEntry": "No entry",
      "legendGreat": "Great",
@@ -3550,7 +3555,8 @@
        "sectionSubtitle": "Composite recovery, stress and strain read from your connected devices",
        "loadError": "Could not load your wellness scores right now.",
        "loadingLabel": "Loading your health scores",
-        "recovery": "Recovery",
+        "recovery": "Recovery score",
+        "recoveryCrossLink": "See your wearable's recovery & strain data",
        "stress": "Stress",
        "strain": "Strain"
      },
@@ -3721,6 +3727,7 @@
    "recovery": {
      "title": "Recovery",
      "description": "How well your body has bounced back — the recovery, strain and load signals your wearable records overnight.",
+      "scoreCrossLink": "See the composite recovery score",
      "emptyTitle": "No recovery data yet",
      "empty": "No recovery data yet. Connect a WHOOP, Polar or Oura account and these scores fill in once your device syncs.",
      "block": {
@@ -4009,7 +4016,7 @@
        },
        "wrapUp": {
          "title": "That’s the tour",
-          "body": "You can revisit any module on its own at any time (“Show this tour” on every page) or restart the whole tour under Settings → Advanced."
+          "body": "This walkthrough runs once — it won't reopen on its own. Everything it covered stays put, and you can turn features on or off any time under Settings → Modules."
        }
      }
    },
@@ -4084,7 +4091,8 @@
      "learning": "Some insights — glucose, sleep rhythm, blood-pressure trends, cycle — sharpen over the first week or two as your data builds up. Empty tiles now just mean we're still learning your baseline.",
      "connectCta": "Connect a device",
      "logCta": "Log your first measurement",
-      "importCta": "Import existing history"
+      "importCta": "Import existing history",
+      "modulesCta": "Choose your modules"
    },
    "welcomeBack": {
      "title": "Welcome back",
@@ -4549,7 +4557,7 @@
        "description": "Structured allergies and family history records."
      },
      "environment": {
-        "title": "Environment",
+        "title": "Location & weather",
        "subtitle": "Weather, daylight, and location context.",
        "disabledHint": "Turn on the Environment module to set a home location and pull in daily weather and daylight.",
        "searchPlaceholder": "Search for a city",
@@ -7959,6 +7967,7 @@
  "mentalHealth": {
    "pageTitle": "Mental wellbeing",
    "pageDescription": "Validated, opt-in self-checks — PHQ-9 (mood), GAD-7 (worry), WHO-5 (well-being) and SCI (sleep) — that sit beside your mood tracking. The value is the trend over time, not any single number.",
+    "moodLink": "View your mood history",
    "attributionLabel": "Attribution",
    "attributionShow": "Source and licence for {instrument}",
    "choosePrompt": "Choose a check-in",
@@ -8404,7 +8413,8 @@
      "noConditions": "Not linked to a condition",
      "previewUnavailable": "No inline preview for this file type",
      "uploadedMeta": "Added {date} · {size}",
-      "share": "Share"
+      "share": "Share",
+      "labValuesLink": "View {count} lab values"
    },
    "dropOverlay": {
      "title": "Drop files to upload",
@@ -8666,7 +8676,9 @@
      "coachCheckin": {
        "title": "Time to check in on a plan",
        "body": "It's been about a week since you set this plan — keep it, adjust it, or let it go. No pressure either way.",
-        "bodyPlan": "About a week ago you set: “{plan}”. Keep it, adjust it, or let it go — no pressure either way."
+        "bodyPlan": "About a week ago you set: “{plan}”. Keep it, adjust it, or let it go — no pressure either way.",
+        "adjustPrompt": "I'd like to adjust my check-in plan.",
+        "adjustPromptPlan": "I'd like to adjust this plan: “{plan}”"
      },
      "tensionWindow": {
        "title": "Possible tension today",
@@ -228,7 +228,7 @@
    "logout": "Cerrar sesión",
    "login": "Iniciar sesión",
    "notifications": "Notificaciones",
-    "theme": "Apariencia",
+    "theme": "Tema",
    "themeSystem": "Sistema",
    "themeDark": "Oscuro",
    "themeLight": "Claro",
@@ -599,12 +599,13 @@
    },
    "watchdog": {
      "title": "Supervisión automática",
-      "description": "HealthLog también vigila en segundo plano — notificaciones de ritmo irregular y alteraciones respiratorias de un reloj conectado, además de alertas de desviación de tu valor de referencia — y solo muestra una tarjeta cuando algo destaca. Que no aparezca ninguna tarjeta es una buena señal, no una función ausente."
+      "description": "HealthLog también vigila en segundo plano — notificaciones de ritmo irregular y alteraciones respiratorias de un reloj conectado, alertas de desviación de tu valor de referencia, y cambios destacables entre tus dos últimos paneles de laboratorio — y solo muestra una tarjeta cuando algo destaca. Que no aparezca ninguna tarjeta es una buena señal, no una función ausente."
    }
  },
  "mood": {
    "title": "Estado de ánimo",
    "subtitle": "Tendencia y registros de estado de ánimo",
+    "mentalWellbeingLink": "Hacer un chequeo de bienestar mental",
    "addEntry": "Añadir",
    "editEntry": "Editar entrada",
    "moodLevel": "Estado de ánimo",
@@ -2263,6 +2264,8 @@
      "historyEmpty": "Tus conversaciones aparecen aquí. Lanza una pregunta para empezar.",
      "historyDeleteAria": "Eliminar conversación",
      "historyDeleted": "Conversación eliminada",
+      "historySearchEmpty": "Ninguna conversación coincide con tu búsqueda.",
+      "historyLoadingMore": "Cargando más…",
      "sourcesTitle": "A qué accedo",
      "sourcesFresh": "Actuales",
      "sourcesStale": "Desactualizados",
@@ -2894,7 +2897,7 @@
      "sleep": {
        "title": "Aún no hay registros de sueño",
        "description": "Una vez disponibles los datos de sueño (manual o vía Withings / Apple Health), aquí aparecerá el desglose por fases.",
-        "cta": "Conectar Withings"
+        "cta": "Conectar una fuente"
      }
    },
    "navWorkouts": "Entrenamientos",
@@ -3262,6 +3265,8 @@
    },
    "navWalkingSpeed": "Velocidad al andar",
    "mood": {
+      "logLink": "Registrar un estado de ánimo",
+      "mentalWellbeingLink": "Hacer un chequeo de bienestar mental",
      "heatmapTitle": "Calendario de ánimo",
      "heatmapNoEntry": "Sin registro",
      "legendGreat": "Genial",
@@ -3550,7 +3555,8 @@
        "sectionSubtitle": "Valores combinados de recuperación, estrés y esfuerzo leídos de tus dispositivos conectados",
        "loadError": "No se pudieron cargar tus puntuaciones de bienestar en este momento.",
        "loadingLabel": "Cargando tus indicadores de salud",
-        "recovery": "Recuperación",
+        "recovery": "Puntuación de recuperación",
+        "recoveryCrossLink": "Ver los datos de recuperación y esfuerzo de tu wearable",
        "stress": "Estrés",
        "strain": "Esfuerzo"
      },
@@ -3721,6 +3727,7 @@
    "recovery": {
      "title": "Recuperación",
      "description": "Cómo se ha recuperado tu cuerpo: las señales de recuperación, esfuerzo y carga que tu wearable registra durante la noche.",
+      "scoreCrossLink": "Ver la puntuación de recuperación combinada",
      "emptyTitle": "Aún no hay datos de recuperación",
      "empty": "Aún no hay datos de recuperación. Conecta una cuenta de WHOOP, Polar u Oura y estas puntuaciones se completarán cuando tu dispositivo sincronice.",
      "block": {
@@ -4009,7 +4016,7 @@
        },
        "wrapUp": {
          "title": "Esto es la visita",
-          "body": "Puedes volver a ver cualquier módulo por separado cuando quieras («Mostrar esta visita» en cada página) o reiniciar la visita completa en Ajustes → Avanzado."
+          "body": "Esta visita guiada se muestra una sola vez, no volverá a abrirse por sí sola. Todo lo que has visto sigue disponible, y puedes activar o desactivar funciones cuando quieras en Ajustes → Módulos."
        }
      }
    },
@@ -4084,7 +4091,8 @@
      "learning": "Algunas conclusiones —glucosa, ritmo de sueño, tendencias de presión arterial, ciclo— se afinan durante la primera semana o dos a medida que se acumulan tus datos. Las casillas vacías ahora solo significan que aún estamos aprendiendo tu línea base.",
      "connectCta": "Conectar un dispositivo",
      "logCta": "Registra tu primera medición",
-      "importCta": "Importar historial existente"
+      "importCta": "Importar historial existente",
+      "modulesCta": "Elige tus módulos"
    },
    "welcomeBack": {
      "title": "Bienvenido de nuevo",
@@ -4549,7 +4557,7 @@
        "description": "Registros estructurados de alergias y antecedentes familiares."
      },
      "environment": {
-        "title": "Entorno",
+        "title": "Ubicación y clima",
        "subtitle": "Contexto de tiempo, luz diurna y ubicación.",
        "disabledHint": "Activa el módulo Entorno para definir una ubicación principal e incorporar el tiempo y las horas de luz diarias.",
        "searchPlaceholder": "Buscar una ciudad",
@@ -7959,6 +7967,7 @@
  "mentalHealth": {
    "pageTitle": "Bienestar mental",
    "pageDescription": "Cuatro autoevaluaciones validadas y opcionales — PHQ-9 (estado de ánimo), GAD-7 (preocupación), WHO-5 (bienestar) y SCI (sueño) — junto a tu registro de ánimo. Lo valioso es la tendencia en el tiempo, no un único número.",
+    "moodLink": "Ver tu historial de estado de ánimo",
    "attributionLabel": "Fuente",
    "attributionShow": "Fuente y licencia de {instrument}",
    "choosePrompt": "Elige una revisión",
@@ -8404,7 +8413,8 @@
      "noConditions": "Sin vincular a una afección",
      "previewUnavailable": "Sin vista previa integrada para este tipo de archivo",
      "uploadedMeta": "Añadido el {date} · {size}",
-      "share": "Compartir"
+      "share": "Compartir",
+      "labValuesLink": "Ver {count} valores de laboratorio"
    },
    "dropOverlay": {
      "title": "Suelta los archivos para subirlos",
@@ -8666,7 +8676,9 @@
      "coachCheckin": {
        "title": "Es buen momento para revisar un plan",
        "body": "Ha pasado alrededor de una semana desde que definiste este plan: consérvalo, ajústalo o déjalo ir. Sin presión.",
-        "bodyPlan": "Hace aproximadamente una semana te propusiste: «{plan}». Consérvalo, ajústalo o déjalo ir, sin presión."
+        "bodyPlan": "Hace aproximadamente una semana te propusiste: «{plan}». Consérvalo, ajústalo o déjalo ir, sin presión.",
+        "adjustPrompt": "Me gustaría ajustar mi plan de seguimiento.",
+        "adjustPromptPlan": "Me gustaría ajustar este plan: «{plan}»"
      },
      "tensionWindow": {
        "title": "Quizá algo de tensión hoy",
@@ -228,7 +228,7 @@
    "logout": "Se déconnecter",
    "login": "Se connecter",
    "notifications": "Notifications",
-    "theme": "Apparence",
+    "theme": "Thème",
    "themeSystem": "Système",
    "themeDark": "Sombre",
    "themeLight": "Clair",
@@ -599,12 +599,13 @@
    },
    "watchdog": {
      "title": "Surveillance automatique",
-      "description": "HealthLog surveille aussi discrètement en arrière-plan — notifications de rythme irrégulier et de perturbations respiratoires envoyées par une montre connectée, ainsi que des alertes d'écart par rapport à ta valeur de référence — et n'affiche une carte que si quelque chose se démarque. L'absence de carte est bon signe, pas une fonctionnalité manquante."
+      "description": "HealthLog surveille aussi discrètement en arrière-plan — notifications de rythme irrégulier et de perturbations respiratoires envoyées par une montre connectée, alertes d'écart par rapport à ta valeur de référence, ainsi que des changements notables entre tes deux derniers bilans de laboratoire — et n'affiche une carte que si quelque chose se démarque. L'absence de carte est bon signe, pas une fonctionnalité manquante."
    }
  },
  "mood": {
    "title": "Humeur",
    "subtitle": "Tendance et entrées d’humeur",
+    "mentalWellbeingLink": "Faire un bilan de bien-être mental",
    "addEntry": "Ajouter",
    "editEntry": "Modifier l’entrée",
    "moodLevel": "Humeur",
@@ -2263,6 +2264,8 @@
      "historyEmpty": "Tes conversations s’affichent ici. Lance une question pour commencer.",
      "historyDeleteAria": "Supprimer la conversation",
      "historyDeleted": "Conversation supprimée",
+      "historySearchEmpty": "Aucune conversation ne correspond à ta recherche.",
+      "historyLoadingMore": "Chargement…",
      "sourcesTitle": "Ce à quoi j’accède",
      "sourcesFresh": "À jour",
      "sourcesStale": "Obsolètes",
@@ -2894,7 +2897,7 @@
      "sleep": {
        "title": "Aucune donnée de sommeil",
        "description": "Dès que des données de sommeil sont disponibles (manuel ou via Withings / Apple Health), la répartition par stade apparaîtra ici.",
-        "cta": "Connecter Withings"
+        "cta": "Connecter une source"
      }
    },
    "navWorkouts": "Entraînements",
@@ -3262,6 +3265,8 @@
    },
    "navWalkingSpeed": "Vitesse de marche",
    "mood": {
+      "logLink": "Enregistrer une humeur",
+      "mentalWellbeingLink": "Faire un bilan de bien-être mental",
      "heatmapTitle": "Calendrier d'humeur",
      "heatmapNoEntry": "Aucune entrée",
      "legendGreat": "Excellent",
@@ -3550,7 +3555,8 @@
        "sectionSubtitle": "Valeurs combinées de récupération, de stress et d'effort issues de tes appareils connectés",
        "loadError": "Impossible de charger vos scores de bien-être pour le moment.",
        "loadingLabel": "Chargement de tes indicateurs de santé",
-        "recovery": "Récupération",
+        "recovery": "Score de récupération",
+        "recoveryCrossLink": "Voir les données de récupération et d'effort de votre wearable",
        "stress": "Stress",
        "strain": "Effort"
      },
@@ -3721,6 +3727,7 @@
    "recovery": {
      "title": "Récupération",
      "description": "À quel point votre corps a récupéré : les signaux de récupération, de charge et d'effort que votre wearable enregistre la nuit.",
+      "scoreCrossLink": "Voir le score de récupération composite",
      "emptyTitle": "Pas encore de données de récupération",
      "empty": "Pas encore de données de récupération. Connectez un compte WHOOP, Polar ou Oura et ces scores se rempliront dès que votre appareil se synchronise.",
      "block": {
@@ -4009,7 +4016,7 @@
        },
        "wrapUp": {
          "title": "C’est la fin de la visite",
-          "body": "Vous pouvez revoir chaque module séparément à tout moment (« Afficher cette visite » sur chaque page) ou relancer la visite complète dans Réglages → Avancé."
+          "body": "Cette visite guidée ne s'affiche qu'une fois — elle ne se rouvrira pas d'elle-même. Tout ce qu'elle a montré reste en place, et vous pouvez activer ou désactiver des fonctionnalités à tout moment dans Réglages → Modules."
        }
      }
    },
@@ -4084,7 +4091,8 @@
      "learning": "Certaines analyses — glycémie, rythme de sommeil, tendances de tension, cycle — s'affinent au cours de la première semaine ou deux à mesure que vos données s'accumulent. Des tuiles vides signifient simplement que nous apprenons encore votre référence.",
      "connectCta": "Connecter un appareil",
      "logCta": "Enregistrez votre première mesure",
-      "importCta": "Importer un historique existant"
+      "importCta": "Importer un historique existant",
+      "modulesCta": "Choisir vos modules"
    },
    "welcomeBack": {
      "title": "Bon retour",
@@ -4549,7 +4557,7 @@
        "description": "Dossiers structurés des allergies et des antécédents familiaux."
      },
      "environment": {
-        "title": "Environnement",
+        "title": "Localisation et météo",
        "subtitle": "Contexte météo, lumière du jour et lieu.",
        "disabledHint": "Activez le module Environnement pour définir un lieu de référence et intégrer la météo et la durée du jour quotidiennes.",
        "searchPlaceholder": "Rechercher une ville",
@@ -7959,6 +7967,7 @@
  "mentalHealth": {
    "pageTitle": "Bien-être mental",
    "pageDescription": "Quatre auto-évaluations validées et facultatives — PHQ-9 (humeur), GAD-7 (inquiétude), WHO-5 (bien-être) et SCI (sommeil) — en complément de ton suivi de l'humeur. Ce qui compte, c'est la tendance dans le temps, pas un chiffre isolé.",
+    "moodLink": "Voir ton historique d'humeur",
    "attributionLabel": "Source",
    "attributionShow": "Source et licence pour {instrument}",
    "choosePrompt": "Choisis un bilan",
@@ -8404,7 +8413,8 @@
      "noConditions": "Lié à aucune affection",
      "previewUnavailable": "Pas d'aperçu intégré pour ce type de fichier",
      "uploadedMeta": "Ajouté le {date} · {size}",
-      "share": "Partager"
+      "share": "Partager",
+      "labValuesLink": "Voir {count} valeurs de laboratoire"
    },
    "dropOverlay": {
      "title": "Déposez les fichiers pour les téléverser",
@@ -8666,7 +8676,9 @@
      "coachCheckin": {
        "title": "Il est temps de revoir un plan",
        "body": "Cela fait environ une semaine que vous avez défini ce plan : gardez-le, ajustez-le ou laissez-le. Sans pression.",
-        "bodyPlan": "Il y a environ une semaine, vous vous étiez fixé : « {plan} ». Gardez-le, ajustez-le ou laissez-le, sans pression."
+        "bodyPlan": "Il y a environ une semaine, vous vous étiez fixé : « {plan} ». Gardez-le, ajustez-le ou laissez-le, sans pression.",
+        "adjustPrompt": "J'aimerais ajuster mon plan de suivi.",
+        "adjustPromptPlan": "J'aimerais ajuster ce plan : « {plan} »"
      },
      "tensionWindow": {
        "title": "Peut-être un peu de tension",
@@ -228,7 +228,7 @@
    "logout": "Esci",
    "login": "Accedi",
    "notifications": "Notifiche",
-    "theme": "Aspetto",
+    "theme": "Tema",
    "themeSystem": "Sistema",
    "themeDark": "Scuro",
    "themeLight": "Chiaro",
@@ -599,12 +599,13 @@
    },
    "watchdog": {
      "title": "Monitoraggio automatico",
-      "description": "HealthLog osserva anche silenziosamente in background — notifiche di ritmo irregolare e disturbi respiratori da un orologio connesso, oltre ad avvisi di scostamento dalla tua base di riferimento — mostrando una scheda solo quando qualcosa emerge. Nessuna scheda visibile è un buon segno, non una funzione mancante."
+      "description": "HealthLog osserva anche silenziosamente in background — notifiche di ritmo irregolare e disturbi respiratori da un orologio connesso, avvisi di scostamento dalla tua base di riferimento, e variazioni degne di nota tra i tuoi due ultimi pannelli di laboratorio — mostrando una scheda solo quando qualcosa emerge. Nessuna scheda visibile è un buon segno, non una funzione mancante."
    }
  },
  "mood": {
    "title": "Umore",
    "subtitle": "Andamento e voci dell’umore",
+    "mentalWellbeingLink": "Fai un check-in sul benessere mentale",
    "addEntry": "Aggiungi",
    "editEntry": "Modifica voce",
    "moodLevel": "Umore",
@@ -2263,6 +2264,8 @@
      "historyEmpty": "Le tue conversazioni compaiono qui. Fai una domanda per iniziare.",
      "historyDeleteAria": "Elimina conversazione",
      "historyDeleted": "Conversazione eliminata",
+      "historySearchEmpty": "Nessuna conversazione corrisponde alla tua ricerca.",
+      "historyLoadingMore": "Caricamento…",
      "sourcesTitle": "A cosa accedo",
      "sourcesFresh": "Aggiornate",
      "sourcesStale": "Datate",
@@ -2894,7 +2897,7 @@
      "sleep": {
        "title": "Nessuna voce di sonno",
        "description": "Quando saranno disponibili dati sul sonno (manuale o tramite Withings / Apple Health), qui apparirà la suddivisione per fase.",
-        "cta": "Collega Withings"
+        "cta": "Collega una fonte"
      }
    },
    "navWorkouts": "Allenamenti",
@@ -3262,6 +3265,8 @@
    },
    "navWalkingSpeed": "Velocità di cammino",
    "mood": {
+      "logLink": "Registra un umore",
+      "mentalWellbeingLink": "Fai un check-in sul benessere mentale",
      "heatmapTitle": "Calendario dell'umore",
      "heatmapNoEntry": "Nessun dato",
      "legendGreat": "Ottimo",
@@ -3550,7 +3555,8 @@
        "sectionSubtitle": "Valori combinati di recupero, stress e sforzo letti dai tuoi dispositivi connessi",
        "loadError": "Al momento non è stato possibile caricare i tuoi punteggi di benessere.",
        "loadingLabel": "Caricamento dei tuoi indicatori di salute",
-        "recovery": "Recupero",
+        "recovery": "Punteggio di recupero",
+        "recoveryCrossLink": "Vedi i dati di recupero e sforzo del tuo wearable",
        "stress": "Stress",
        "strain": "Sforzo"
      },
@@ -3721,6 +3727,7 @@
    "recovery": {
      "title": "Recupero",
      "description": "Quanto bene il tuo corpo si è ripreso: i segnali di recupero, sforzo e carico che il tuo wearable registra durante la notte.",
+      "scoreCrossLink": "Vedi il punteggio di recupero composito",
      "emptyTitle": "Ancora nessun dato di recupero",
      "empty": "Ancora nessun dato di recupero. Collega un account WHOOP, Polar o Oura e questi punteggi si popoleranno appena il dispositivo si sincronizza.",
      "block": {
@@ -4009,7 +4016,7 @@
        },
        "wrapUp": {
          "title": "Questa è la visita guidata",
-          "body": "Puoi rivedere ogni modulo singolarmente in qualsiasi momento («Mostra questa visita» su ogni pagina) o riavviare l’intera visita da Impostazioni → Avanzate."
+          "body": "Questa visita guidata viene mostrata una sola volta: non si riaprirà da sola. Tutto ciò che hai visto resta al suo posto e puoi attivare o disattivare le funzioni quando vuoi da Impostazioni → Moduli."
        }
      }
    },
@@ -4084,7 +4091,8 @@
      "learning": "Alcune analisi — glicemia, ritmo del sonno, tendenze della pressione, ciclo — si affinano nella prima settimana o due man mano che i tuoi dati si accumulano. Le caselle vuote ora indicano solo che stiamo ancora imparando il tuo riferimento.",
      "connectCta": "Collega un dispositivo",
      "logCta": "Registra la tua prima misurazione",
-      "importCta": "Importa la cronologia esistente"
+      "importCta": "Importa la cronologia esistente",
+      "modulesCta": "Scegli i tuoi moduli"
    },
    "welcomeBack": {
      "title": "Bentornato",
@@ -4549,7 +4557,7 @@
        "description": "Record strutturati di allergie e anamnesi familiare."
      },
      "environment": {
-        "title": "Ambiente",
+        "title": "Posizione e meteo",
        "subtitle": "Contesto meteo, luce del giorno e posizione.",
        "disabledHint": "Attiva il modulo Ambiente per impostare una posizione principale e integrare il meteo e le ore di luce giornaliere.",
        "searchPlaceholder": "Cerca una città",
@@ -7959,6 +7967,7 @@
  "mentalHealth": {
    "pageTitle": "Benessere mentale",
    "pageDescription": "Quattro autovalutazioni validate e facoltative — PHQ-9 (umore), GAD-7 (preoccupazione), WHO-5 (benessere) e SCI (sonno) — accanto al tuo diario dell'umore. Ciò che conta è l'andamento nel tempo, non un singolo numero.",
+    "moodLink": "Vedi il tuo andamento dell'umore",
    "attributionLabel": "Fonte",
    "attributionShow": "Fonte e licenza di {instrument}",
    "choosePrompt": "Scegli un check-in",
@@ -8404,7 +8413,8 @@
      "noConditions": "Non collegato a una patologia",
      "previewUnavailable": "Nessuna anteprima integrata per questo tipo di file",
      "uploadedMeta": "Aggiunto il {date} · {size}",
-      "share": "Condividi"
+      "share": "Condividi",
+      "labValuesLink": "Vedi {count} valori di laboratorio"
    },
    "dropOverlay": {
      "title": "Rilascia i file per caricarli",
@@ -8666,7 +8676,9 @@
      "coachCheckin": {
        "title": "È il momento di rivedere un piano",
        "body": "È passata circa una settimana da quando hai definito questo piano: mantienilo, modificalo o lascialo andare. Senza pressioni.",
-        "bodyPlan": "Circa una settimana fa ti eri proposto: «{plan}». Mantienilo, modificalo o lascialo andare, senza pressioni."
+        "bodyPlan": "Circa una settimana fa ti eri proposto: «{plan}». Mantienilo, modificalo o lascialo andare, senza pressioni.",
+        "adjustPrompt": "Vorrei modificare il mio piano di controllo.",
+        "adjustPromptPlan": "Vorrei modificare questo piano: «{plan}»"
      },
      "tensionWindow": {
        "title": "Forse un po' di tensione oggi",
@@ -228,7 +228,7 @@
    "logout": "Wyloguj",
    "login": "Zaloguj",
    "notifications": "Powiadomienia",
-    "theme": "Wygląd",
+    "theme": "Motyw",
    "themeSystem": "Systemowy",
    "themeDark": "Ciemny",
    "themeLight": "Jasny",
@@ -599,12 +599,13 @@
    },
    "watchdog": {
      "title": "Automatyczny nadzór",
-      "description": "HealthLog dyskretnie obserwuje też dane w tle — powiadomienia o nieregularnym rytmie i zaburzeniach oddychania z podłączonego zegarka, a także alerty o odchyleniu od Twojej wartości bazowej — i pokazuje kartę tylko wtedy, gdy coś zwraca uwagę. Brak karty to dobry znak, a nie brakująca funkcja."
+      "description": "HealthLog dyskretnie obserwuje też dane w tle — powiadomienia o nieregularnym rytmie i zaburzeniach oddychania z podłączonego zegarka, alerty o odchyleniu od Twojej wartości bazowej oraz zauważalne zmiany między dwoma ostatnimi panelami badań laboratoryjnych — i pokazuje kartę tylko wtedy, gdy coś zwraca uwagę. Brak karty to dobry znak, a nie brakująca funkcja."
    }
  },
  "mood": {
    "title": "Nastrój",
    "subtitle": "Trend i wpisy nastroju",
+    "mentalWellbeingLink": "Wykonaj sprawdzian dobrostanu psychicznego",
    "addEntry": "Dodaj",
    "editEntry": "Edytuj wpis",
    "moodLevel": "Nastrój",
@@ -2263,6 +2264,8 @@
      "historyEmpty": "Twoje rozmowy pojawią się tutaj. Zadaj pytanie, aby zacząć.",
      "historyDeleteAria": "Usuń rozmowę",
      "historyDeleted": "Rozmowa usunięta",
+      "historySearchEmpty": "Żadna rozmowa nie pasuje do wyszukiwania.",
+      "historyLoadingMore": "Wczytywanie…",
      "sourcesTitle": "Do czego mam dostęp",
      "sourcesFresh": "Aktualne",
      "sourcesStale": "Nieaktualne",
@@ -2894,7 +2897,7 @@
      "sleep": {
        "title": "Brak danych o śnie",
        "description": "Gdy pojawią się dane o śnie (ręcznie lub przez Withings / Apple Health), wyświetli się tu rozkład faz na noc.",
-        "cta": "Połącz Withings"
+        "cta": "Połącz źródło"
      }
    },
    "navWorkouts": "Treningi",
@@ -3262,6 +3265,8 @@
    },
    "navWalkingSpeed": "Prędkość chodu",
    "mood": {
+      "logLink": "Zapisz nastrój",
+      "mentalWellbeingLink": "Wykonaj sprawdzian dobrostanu psychicznego",
      "heatmapTitle": "Kalendarz nastroju",
      "heatmapNoEntry": "Brak wpisu",
      "legendGreat": "Świetnie",
@@ -3550,7 +3555,8 @@
        "sectionSubtitle": "Złożone wartości regeneracji, stresu i obciążenia odczytane z Twoich połączonych urządzeń",
        "loadError": "Nie udało się teraz wczytać Twoich wyników dobrostanu.",
        "loadingLabel": "Wczytywanie Twoich wskaźników zdrowia",
-        "recovery": "Regeneracja",
+        "recovery": "Wynik regeneracji",
+        "recoveryCrossLink": "Zobacz dane regeneracji i obciążenia z Twojego urządzenia",
        "stress": "Stres",
        "strain": "Obciążenie"
      },
@@ -3721,6 +3727,7 @@
    "recovery": {
      "title": "Regeneracja",
      "description": "Jak dobrze Twój organizm się zregenerował – sygnały regeneracji, obciążenia i load, które Twoje urządzenie rejestruje w nocy.",
+      "scoreCrossLink": "Zobacz łączny wynik regeneracji",
      "emptyTitle": "Brak danych o regeneracji",
      "empty": "Brak danych o regeneracji. Połącz konto WHOOP, Polar lub Oura, a te wyniki uzupełnią się po synchronizacji urządzenia.",
      "block": {
@@ -4009,7 +4016,7 @@
        },
        "wrapUp": {
          "title": "To koniec przewodnika",
-          "body": "Każdy moduł możesz obejrzeć osobno w dowolnej chwili („Pokaż ten przewodnik” na każdej stronie) lub uruchomić cały przewodnik ponownie w Ustawienia → Zaawansowane."
+          "body": "Ten przewodnik pokazuje się tylko raz — nie otworzy się ponownie sam. Wszystko, co pokazał, zostaje na miejscu, a funkcje możesz włączać i wyłączać w dowolnym momencie w Ustawienia → Moduły."
        }
      }
    },
@@ -4084,7 +4091,8 @@
      "learning": "Niektóre wnioski — glukoza, rytm snu, trendy ciśnienia, cykl — wyostrzają się w ciągu pierwszego tygodnia lub dwóch, w miarę gromadzenia danych. Puste kafelki oznaczają teraz tylko, że wciąż poznajemy Twój punkt odniesienia.",
      "connectCta": "Połącz urządzenie",
      "logCta": "Zapisz pierwszy pomiar",
-      "importCta": "Importuj istniejącą historię"
+      "importCta": "Importuj istniejącą historię",
+      "modulesCta": "Wybierz swoje moduły"
    },
    "welcomeBack": {
      "title": "Witaj ponownie",
@@ -4549,7 +4557,7 @@
        "description": "Ustrukturyzowane rekordy alergii i wywiadu rodzinnego."
      },
      "environment": {
-        "title": "Środowisko",
+        "title": "Lokalizacja i pogoda",
        "subtitle": "Kontekst pogody, światła dziennego i lokalizacji.",
        "disabledHint": "Włącz moduł Środowisko, aby ustawić lokalizację domową i pobierać dzienną pogodę oraz długość dnia.",
        "searchPlaceholder": "Szukaj miasta",
@@ -7959,6 +7967,7 @@
  "mentalHealth": {
    "pageTitle": "Dobrostan psychiczny",
    "pageDescription": "Cztery zwalidowane, dobrowolne testy — PHQ-9 (nastrój), GAD-7 (lęk), WHO-5 (samopoczucie) i SCI (sen) — obok Twojego dziennika nastroju. Liczy się trend w czasie, a nie pojedynczy wynik.",
+    "moodLink": "Zobacz historię swojego nastroju",
    "attributionLabel": "Źródło",
    "attributionShow": "Źródło i licencja: {instrument}",
    "choosePrompt": "Wybierz sprawdzenie",
@@ -8404,7 +8413,8 @@
      "noConditions": "Niepowiązany ze schorzeniem",
      "previewUnavailable": "Brak podglądu dla tego typu pliku",
      "uploadedMeta": "Dodano {date} · {size}",
-      "share": "Udostępnij"
+      "share": "Udostępnij",
+      "labValuesLink": "Zobacz {count} wyników laboratoryjnych"
    },
    "dropOverlay": {
      "title": "Upuść pliki, aby je przesłać",
@@ -8666,7 +8676,9 @@
      "coachCheckin": {
        "title": "Czas przyjrzeć się planowi",
        "body": "Minął mniej więcej tydzień, odkąd ustaliłeś ten plan — zachowaj go, dostosuj lub odpuść. Bez presji.",
-        "bodyPlan": "Około tygodnia temu postanowiłeś: „{plan}”. Zachowaj to, dostosuj lub odpuść — bez presji."
+        "bodyPlan": "Około tygodnia temu postanowiłeś: „{plan}”. Zachowaj to, dostosuj lub odpuść — bez presji.",
+        "adjustPrompt": "Chciałbym dostosować mój plan kontrolny.",
+        "adjustPromptPlan": "Chciałbym dostosować ten plan: „{plan}”"
      },
      "tensionWindow": {
        "title": "Może chwila napięcia dzisiaj",
@@ -1,6 +1,6 @@
 {
  "name": "healthlog",
-  "version": "1.30.1",
+  "version": "1.30.2",
  "description": "Self-hosted personal-health-tracking PWA with Withings integration, AI insights, and doctor-report PDF export.",
  "license": "PolyForm-Noncommercial-1.0.0",
  "homepage": "https://healthlog.dev",
@@ -36,7 +36,7 @@ try {
 // v1.4.38.4 → v1.4.42. Do not hand-edit; bump `package.json` and rebuild.
 const CACHE_VERSION =
  (typeof self !== "undefined" && self.__APP_VERSION__) ||
-  /* @sw-version-fallback */ "v1.30.1";
+  /* @sw-version-fallback */ "v1.30.2";
 const STATIC_CACHE = `healthlog-static-${CACHE_VERSION}`;
 const PAGE_CACHE = `healthlog-pages-${CACHE_VERSION}`;
 // v1.18.6 — read-only data cache for a curated allowlist of safe GET `/api/*`
@@ -0,0 +1,176 @@
+import { describe, it, expect, vi, beforeEach } from "vitest";
+
+/**
+ * v1.30.2 (QoL H1) — `GET /api/insights/chat` gained an optional `q` search
+ * param so the history rail's search reaches the caller's FULL conversation
+ * set (title-only, server-side) instead of filtering only the loaded page.
+ * This suite pins the route's parsing: trimmed, length-capped, and omitted
+ * (not an empty string) when absent — `listConversations` treats `undefined`
+ * and `""` identically, but keeping the route from ever forwarding a bare
+ * empty string documents the contract at the call site.
+ */
+
+vi.mock("@/lib/api-handler", () => ({
+  apiHandler: <T extends (...args: unknown[]) => unknown>(fn: T) => fn,
+  requireAuth: vi.fn(async () => ({ user: { id: "u1", locale: "en" } })),
+  HttpError: class HttpError extends Error {
+    status: number;
+    constructor(status: number, message: string) {
+      super(message);
+      this.status = status;
+    }
+  },
+}));
+
+const { requireModuleEnabled, requireAssistantSurface } = vi.hoisted(() => ({
+  requireModuleEnabled: vi.fn(async () => ({ enabled: true })),
+  requireAssistantSurface: vi.fn(async () => undefined),
+}));
+vi.mock("@/lib/modules/gate", () => ({ requireModuleEnabled }));
+vi.mock("@/lib/feature-flags", () => ({ requireAssistantSurface }));
+
+vi.mock("@/lib/api-response", () => ({
+  apiError: (error: string, status: number, meta?: unknown) => ({
+    data: null,
+    error,
+    status,
+    meta,
+  }),
+  apiSuccess: (data: unknown) => ({ data, error: null, status: 200 }),
+}));
+vi.mock("@/lib/logging/context", () => ({ annotate: vi.fn() }));
+vi.mock("@/lib/auth/audit", () => ({ auditLog: vi.fn() }));
+vi.mock("@/lib/db", () => ({ prisma: {} }));
+vi.mock("@/lib/rate-limit", () => ({ checkRateLimit: vi.fn() }));
+vi.mock("@/lib/i18n/server-locale", () => ({
+  resolveServerLocale: vi.fn(async () => "en"),
+}));
+vi.mock("@/lib/ai/provider-runner", () => ({
+  AllProvidersFailedError: class extends Error {},
+  runRawCompletionWithFallback: vi.fn(),
+  runStreamingRawCompletionWithFallback: vi.fn(),
+}));
+vi.mock("@/lib/ai/provider", () => ({
+  resolveProviderChain: vi.fn(),
+  resolveProvider: vi.fn(),
+}));
+vi.mock("@/lib/ai/consent-guard", () => ({ assertConsentForChain: vi.fn() }));
+vi.mock("@/lib/ai/prompts/insight-generator", () => ({ PROMPT_VERSION: "x" }));
+vi.mock("@/lib/ai/coach/types", () => ({
+  coachChatRequestSchema: { safeParse: () => ({ success: false }) },
+}));
+vi.mock("@/lib/ai/coach/tools", () => ({
+  COACH_TOOL_DEFS: [],
+  MAX_ROUNDS: 3,
+  buildCoachDataInventory: vi.fn(),
+  renderDataInventory: vi.fn(),
+  renderFocusHint: vi.fn(() => ""),
+  buildToolModeAddendum: vi.fn(),
+  runCoachToolLoop: vi.fn(),
+}));
+
+interface ListConversationsCallArgs {
+  userId: string;
+  cursor?: string | null;
+  limit?: number;
+  q?: string;
+}
+
+// `vi.mock` factories are hoisted above every top-level declaration in the
+// file, so a plain `const listConversations = vi.fn(...)` referenced inside
+// the factory below would hit the temporal dead zone. `vi.hoisted` runs
+// before that hoisting and is the documented escape hatch.
+const { listConversations } = vi.hoisted(() => ({
+  listConversations: vi.fn(async (params: ListConversationsCallArgs) => {
+    void params; // typed purely so `.mock.calls[n][0]` below is not `never`
+    return { conversations: [], nextCursor: null };
+  }),
+}));
+vi.mock("@/lib/ai/coach/persistence", () => ({
+  appendMessage: vi.fn(),
+  createConversation: vi.fn(),
+  fetchConversationWithMessages: vi.fn(),
+  listConversations,
+}));
+vi.mock("@/lib/ai/coach/coach-memory-shared", () => ({
+  enqueueCoachMemoryRefresh: vi.fn(),
+}));
+vi.mock("@/lib/ai/coach/facts", () => ({ storeDeterministicFacts: vi.fn() }));
+vi.mock("@/lib/ai/coach/budget", () => ({
+  buildDateKey: vi.fn(),
+  reserveBudget: vi.fn(),
+  reconcileSpend: vi.fn(),
+  resolveDailyCap: vi.fn(() => 200_000),
+}));
+vi.mock("@/lib/ai/coach/refusal", () => ({ detectRefusal: vi.fn() }));
+vi.mock("@/lib/ai/coach/system-prompt", () => ({
+  getCoachSystemPrompt: vi.fn(),
+}));
+vi.mock("@/lib/ai/coach/about-me", () => ({
+  getSelfContextTextForUser: vi.fn(),
+}));
+vi.mock("@/lib/ai/coach/snapshot", () => ({ buildCoachSnapshot: vi.fn() }));
+vi.mock("@/lib/ai/coach/keyvalues", () => ({
+  parseKeyValuesSentinel: vi.fn(),
+}));
+vi.mock("@/lib/validations/coach-prefs", () => ({ parseCoachPrefs: vi.fn() }));
+vi.mock("@/lib/sse/create-stream", () => ({ createSseStream: vi.fn() }));
+
+import { GET } from "../route";
+
+type Envelope = { data: unknown; error: string | null; status: number };
+const get = GET as unknown as (req: Request) => Promise<Envelope>;
+
+function listReq(qs = ""): Request {
+  return new Request(`http://localhost/api/insights/chat${qs}`, {
+    method: "GET",
+  });
+}
+
+describe("GET /api/insights/chat — q search param", () => {
+  beforeEach(() => {
+    vi.clearAllMocks();
+    requireModuleEnabled.mockResolvedValue({ enabled: true });
+    requireAssistantSurface.mockResolvedValue(undefined);
+    listConversations.mockResolvedValue({
+      conversations: [],
+      nextCursor: null,
+    });
+  });
+
+  it("omits q from listConversations when absent", async () => {
+    await get(listReq());
+    expect(listConversations).toHaveBeenCalledWith(
+      expect.objectContaining({ userId: "u1", q: undefined }),
+    );
+  });
+
+  it("trims and forwards q", async () => {
+    await get(listReq("?q=%20blood%20pressure%20"));
+    expect(listConversations).toHaveBeenCalledWith(
+      expect.objectContaining({ q: "blood pressure" }),
+    );
+  });
+
+  it("caps an oversized q at 200 chars", async () => {
+    const long = "a".repeat(500);
+    await get(listReq(`?q=${long}`));
+    const call = listConversations.mock.calls[0][0];
+    expect(call.q).toHaveLength(200);
+  });
+
+  it("treats a whitespace-only q the same as absent (undefined, not empty string)", async () => {
+    await get(listReq("?q=%20%20%20"));
+    const call = listConversations.mock.calls[0][0];
+    // Trimmed to "" — the route still passes the (falsy) trimmed string
+    // through; listConversations treats "" the same as undefined.
+    expect(call.q).toBe("");
+  });
+
+  it("still forwards cursor + limit unchanged alongside q", async () => {
+    await get(listReq("?cursor=abc&limit=10&q=weight"));
+    expect(listConversations).toHaveBeenCalledWith(
+      expect.objectContaining({ cursor: "abc", limit: 10, q: "weight" }),
+    );
+  });
+});
@@ -1435,13 +1435,25 @@ function streamProviderError(args: { code: string }): Response {
 // the conversationId existence check + 20-turn cap.
 export const POST = apiHandler(handleChatRequest);

+/** v1.30.2 (QoL H1) — hard cap on the `?q=` search string length. */
+const LIST_QUERY_MAX_LEN = 200;
+
 /**
- * GET /api/insights/chat?cursor=<id>&limit=<n>
+ * GET /api/insights/chat?cursor=<id>&limit=<n>&q=<text>
 *
 * Cursor-paginated list of the caller's conversations for the rail.
 * Default limit 20, hard cap 50. Cursor is the id of the last item
 * on the previous page; callers receive `{ nextCursor: null }` when
 * they have reached the end.
+ *
+ * v1.30.2 (QoL H1) — optional `q` narrows the page to conversations whose
+ * TITLE contains the text (case-insensitive substring). This makes the
+ * history rail's search reach the caller's FULL conversation set instead
+ * of only the loaded page; the client re-issues the cursor walk under the
+ * new `q` whenever the search box changes rather than filtering client-
+ * side. Message bodies are encrypted at rest and are NOT searched — a
+ * decrypt-and-scan over every message would be prohibitively expensive
+ * for a live keystroke search and is out of scope for this pass.
 */
 export const GET = apiHandler(async (request: NextRequest) => {
  const auth = await requireAuth();
@@ -1456,11 +1468,14 @@ export const GET = apiHandler(async (request: NextRequest) => {
  const cursor = url.searchParams.get("cursor");
  const limitRaw = url.searchParams.get("limit");
  const limit = limitRaw ? Number.parseInt(limitRaw, 10) : undefined;
+  const qRaw = url.searchParams.get("q");
+  const q = qRaw ? qRaw.trim().slice(0, LIST_QUERY_MAX_LEN) : undefined;

  const page = await listConversations({
    userId: auth.user.id,
    cursor,
    limit: Number.isFinite(limit) ? (limit as number) : undefined,
+    q,
    // v1.28.51 (Documents R3, Design A) — the rail now surfaces BOTH health
    // threads and doc-scoped threads (the DTO carries `documentId` +
    // `documentTitle` so the client badges the fenced ones). Omitting the
@@ -1472,7 +1487,7 @@ export const GET = apiHandler(async (request: NextRequest) => {

  annotate({
    action: { name: "insights.coach.list" },
-    meta: { count: page.conversations.length },
+    meta: { count: page.conversations.length, hasQuery: Boolean(q) },
  });

  return apiSuccess(page);
@@ -196,6 +196,27 @@ describe("GET /api/insights/correlations", () => {
    expect(prisma.measurement.findMany).not.toHaveBeenCalled();
  });

+  // v1.30 (DATAINT M1) — `asc` + `take` truncated the RECENT end of the
+  // 180-day window on a dense account; the read must order `desc` so the
+  // cap keeps the newest rows (re-sorted `asc` in JS before series-building).
+  it("reads the measurement + mood window ordered desc so the cap keeps the newest rows", async () => {
+    vi.mocked(getSession).mockResolvedValue(SESSION_OK as never);
+    await callGet(makeReq());
+
+    expect(prisma.measurement.findMany).toHaveBeenCalledWith(
+      expect.objectContaining({
+        orderBy: { measuredAt: "desc" },
+        take: 20000,
+      }),
+    );
+    expect(prisma.moodEntry.findMany).toHaveBeenCalledWith(
+      expect.objectContaining({
+        orderBy: { moodLoggedAt: "desc" },
+        take: 5000,
+      }),
+    );
+  });
+
  // v1.18.0 (B2) — the route now also requires the `insights` module.
  it("returns 403 + module.disabled when the insights module is off", async () => {
    vi.mocked(getSession).mockResolvedValue(SESSION_OK as never);
@@ -101,7 +101,17 @@ export const GET = apiHandler(async () => {
  // `buildMeasurementDailySeries`'s per-type grain resolution below (source
  // collapse for cumulative types, per-night reconstruction for sleep); the
  // rest of the channels never look at them.
-  const [measurements, moodEntries, priorityJson] = await Promise.all([
+  // v1.30 (PERFAUDIT M1) — `orderBy asc` + `take` truncates from the START
+  // of the window. A dense account (unconsolidated step chunks alone can
+  // run ~200 rows/day → 36 000 rows over 180 days) can exceed the cap, and
+  // an ascending order then drops the NEWEST rows — exactly backwards for
+  // `discoverEmergingCorrelations`, whose entire job is the recent window.
+  // Order DESC so the cap keeps the most-recent rows, then re-sort ASC in
+  // JS before folding into per-type series (every downstream day-bucketing
+  // pass is order-insensitive, but callers document an ascending contract).
+  const MEASUREMENT_READ_CAP = 20000;
+  const MOOD_READ_CAP = 5000;
+  const [measurementsDesc, moodEntriesDesc, priorityJson] = await Promise.all([
    prisma.measurement.findMany({
      where: {
        userId: user.id,
@@ -109,8 +119,8 @@ export const GET = apiHandler(async () => {
        measuredAt: { gte: since },
        type: { in: [...behaviourTypes, ...outcomeTypes] },
      },
-      orderBy: { measuredAt: "asc" },
-      take: 20000,
+      orderBy: { measuredAt: "desc" },
+      take: MEASUREMENT_READ_CAP,
      select: {
        type: true,
        value: true,
@@ -122,12 +132,20 @@ export const GET = apiHandler(async () => {
    }),
    prisma.moodEntry.findMany({
      where: { userId: user.id, deletedAt: null, moodLoggedAt: { gte: since } },
-      orderBy: { moodLoggedAt: "asc" },
-      take: 5000,
+      orderBy: { moodLoggedAt: "desc" },
+      take: MOOD_READ_CAP,
      select: { score: true, moodLoggedAt: true },
    }),
    loadUserSourcePriority(user.id),
  ]);
+  const measurementsCapped = measurementsDesc.length >= MEASUREMENT_READ_CAP;
+  const moodCapped = moodEntriesDesc.length >= MOOD_READ_CAP;
+  const measurements = [...measurementsDesc].sort(
+    (a, b) => a.measuredAt.getTime() - b.measuredAt.getTime(),
+  );
+  const moodEntries = [...moodEntriesDesc].sort(
+    (a, b) => a.moodLoggedAt.getTime() - b.moodLoggedAt.getTime(),
+  );

  const measurementsByType = new Map<string, MeasurementSeriesRow[]>();
  for (const m of measurements) {
@@ -221,6 +239,13 @@ export const GET = apiHandler(async () => {
      pairs_tested: result.pairsTested,
      discovered: result.discovered.length,
      fdr_q: result.fdrQ,
+      // PERFAUDIT M1 — surfaces when a dense account's window exceeded the
+      // read cap. The cap now falls on the OLDEST rows (desc + take), so a
+      // capped read still covers the recent window `discoverEmergingCorrelations`
+      // needs; this only tells a dashboard the retrospective scan's older
+      // half of the window may be thin.
+      measurements_capped: measurementsCapped,
+      mood_entries_capped: moodCapped,
      // FDREXTEND — per-channel day-counts so a dashboard can see whether the
      // two sparse new channels reached the n ≥ 20 floor or degraded to absent.
      compliance_days: complianceSeries.points.length,
@@ -132,6 +132,32 @@ describe("GET /api/nutrients", () => {
    ]);
  });

+  it("counts a multi-source day once (DATAINT M3 regression)", async () => {
+    // v1.29 migration 0249 lets a day carry one row per source. A non-latest
+    // day logged via both APPLE_HEALTH and MANUAL must still count as ONE
+    // day, not two.
+    vi.mocked(prisma.nutrientIntakeDay.findMany).mockResolvedValue([
+      { nutrient: "water", unit: "ml", day: "2026-07-13", amount: 900 },
+      { nutrient: "water", unit: "ml", day: "2026-07-12", amount: 600 },
+      { nutrient: "water", unit: "ml", day: "2026-07-12", amount: 400 },
+      { nutrient: "water", unit: "ml", day: "2026-07-11", amount: 500 },
+    ] as never);
+
+    const res = await GET(getReq("?days=7"));
+    const body = (await res.json()) as OverviewResponse;
+    const water = body.data.nutrients.find((n) => n.nutrient === "water");
+    expect(water).toEqual({
+      nutrient: "water",
+      unit: "ml",
+      latestDay: "2026-07-13",
+      latestAmount: 900,
+      // Three distinct days (13th, 12th, 11th) — the buggy version counted
+      // four, since the 12th's second source row didn't match the pinned
+      // `latestDay` (the 13th) and opened a phantom extra day.
+      daysWithData: 3,
+    });
+  });
+
  it("scopes the query to the authenticated user and the window floor", async () => {
    await GET(getReq("?days=14"));
    expect(prisma.nutrientIntakeDay.findMany).toHaveBeenCalledWith(
@@ -17,6 +17,7 @@ import { annotate } from "@/lib/logging/context";
 import { requireModuleEnabled } from "@/lib/modules/gate";
 import { NUTRIENT_CODES, isNutrientCode } from "@/lib/nutrients/catalog";
 import { nutrientOverviewQuerySchema } from "@/lib/validations/nutrients";
+import { DEFAULT_TIMEZONE, shiftDateKey, userDayKey } from "@/lib/tz/format";

 export const dynamic = "force-dynamic";

@@ -36,12 +37,14 @@ export const GET = apiHandler(async (request: NextRequest) => {
  }
  const { days } = parsed.data;

-  // Window floor as a UTC day key. The stored `day` is a local-timezone
-  // key, so the UTC floor is off by at most one calendar day at the
-  // window edge — fine for a "last N days" summary card.
-  const since = new Date(Date.now() - (days - 1) * 24 * 60 * 60 * 1000)
-    .toISOString()
-    .slice(0, 10);
+  // v1.30 (DATAINT L4) — window floor anchored on the caller's own local
+  // "today" (`User.timezone`), mirroring the sibling `GET
+  // /api/nutrients/daily` route. The stored `day` is a local-timezone key,
+  // so a UTC floor was off by up to a calendar day at the window edge for
+  // any non-UTC user.
+  const userTz = user.timezone || DEFAULT_TIMEZONE;
+  const todayKey = userDayKey(new Date(), userTz);
+  const since = shiftDateKey(todayKey, -(days - 1));

  const rows = await prisma.nutrientIntakeDay.findMany({
    where: { userId: user.id, day: { gte: since } },
@@ -50,14 +53,26 @@ export const GET = apiHandler(async (request: NextRequest) => {
  });

  // v1.29 — `source` joined the PK (migration 0249): a day can now carry
-  // an APPLE_HEALTH row AND a MANUAL row, so the fold sums amounts
-  // WITHIN (nutrient, day) before folding across days. Rows arrive
-  // day-DESC inside each nutrient, so the first DAY seen per code is
-  // the latest; a second row for that same day (the other source) adds
-  // to the running total instead of opening a new "day".
+  // an APPLE_HEALTH row AND a MANUAL row. Rows arrive day-DESC inside each
+  // nutrient, so the first DAY seen per code is the latest; a second row
+  // for that same day (the other source) adds to the running total instead
+  // of opening a new "day".
+  //
+  // v1.30 (DATAINT M3) — `daysWithData` used to increment on every row
+  // whose `day` differed from the PINNED `latestDay` field, so a second
+  // source's row on an already-counted OLDER day (e.g. water logged via
+  // both APPLE_HEALTH and MANUAL on the same non-latest day) opened a
+  // second "day" for that same calendar date. `daysSeen` is a per-nutrient
+  // set of distinct day keys — every row adds to it, so a repeated day
+  // (any number of source rows) is counted exactly once.
  const byCode = new Map<
    string,
-    { unit: string; latestDay: string; latestAmount: number; days: number }
+    {
+      unit: string;
+      latestDay: string;
+      latestAmount: number;
+      daysSeen: Set<string>;
+    }
  >();
  for (const row of rows) {
    const existing = byCode.get(row.nutrient);
@@ -66,12 +81,13 @@ export const GET = apiHandler(async (request: NextRequest) => {
        unit: row.unit,
        latestDay: row.day,
        latestAmount: row.amount,
-        days: 1,
+        daysSeen: new Set([row.day]),
      });
-    } else if (row.day === existing.latestDay) {
+      continue;
+    }
+    existing.daysSeen.add(row.day);
+    if (row.day === existing.latestDay) {
      existing.latestAmount += row.amount;
-    } else {
-      existing.days += 1;
    }
  }

@@ -86,7 +102,7 @@ export const GET = apiHandler(async (request: NextRequest) => {
      unit: summary.unit,
      latestDay: summary.latestDay,
      latestAmount: summary.latestAmount,
-      daysWithData: summary.days,
+      daysWithData: summary.daysSeen.size,
    };
  });

@@ -30,10 +30,18 @@
 *   slice the canonical projection. A naive `skip` + per-window dedup
 *   double-counts boundary clusters across pages because a cluster
 *   whose first member sits in page N-1 and whose later member sits in
- *   page N would be dropped on N-1 and surface again on N. Workout
- *   volume for HealthLog is bounded (single user, tens of workouts per
- *   week — single-digit MB at the table's busiest), so reading the
- *   full filtered set per request is well within budget.
+ *   page N would be dropped on N-1 and surface again on N.
+ *
+ *   The canonical projection (the deduped, sorted row set) is cached
+ *   keyed on the FILTER (`userId|since|until|sportType`) — deliberately
+ *   WITHOUT `offset`/`limit` — so every page of a drain shares one cache
+ *   entry: the first page's request builds it (one `findMany` + one
+ *   dedup pass), every later page in the same TTL window slices the
+ *   already-built array for free. Before this, each page carried its own
+ *   cache key, so a full historical drain (the iOS app's "drain
+ *   historical workouts" path, potentially thousands of rows on a
+ *   10-year Apple Health import) re-ran the full read + dedup once per
+ *   page — O(pages × rows) instead of O(rows).
 *
 * Response shape (iOS contract):
 *   - `workouts: WorkoutListEntry[]` — each entry exposes
@@ -68,6 +76,32 @@ interface WorkoutsParams {
  readonly sportType: string | null;
 }

+type WorkoutFilterParams = Pick<
+  WorkoutsParams,
+  "since" | "until" | "sportType"
+>;
+
+interface CanonicalWorkoutRow {
+  id: string;
+  source: string;
+  externalId: string | null;
+  sportType: string | null;
+  startedAt: Date | null;
+  endedAt: Date | null;
+  durationSec: number | null;
+  totalDistanceM: number | null;
+  totalEnergyKcal: number | null;
+  avgHeartRate: number | null;
+  maxHeartRate: number | null;
+  createdAt: Date;
+}
+
+/** The cached, deduped, most-recent-first row set for one filter combination. */
+interface WorkoutsProjection {
+  canonical: CanonicalWorkoutRow[];
+  rawCount: number;
+}
+
 interface WorkoutsResult {
  workouts: Array<{
    id: string;
@@ -90,11 +124,11 @@ interface WorkoutsResult {
  };
 }

-async function buildWorkoutsResponse(
+async function buildWorkoutsProjection(
  userId: string,
-  params: WorkoutsParams,
-): Promise<WorkoutsResult> {
-  const { limit, offset, since, until, sportType } = params;
+  params: WorkoutFilterParams,
+): Promise<WorkoutsProjection> {
+  const { since, until, sportType } = params;

  const where: Record<string, unknown> = { userId };
  if (since || until) {
@@ -115,29 +149,34 @@ async function buildWorkoutsResponse(
    where.sportType = sportType;
  }

-  const rows = await prisma.workout.findMany({
-    where,
-    orderBy: [{ startedAt: "desc" }, { id: "asc" }],
-    select: {
-      id: true,
-      source: true,
-      externalId: true,
-      sportType: true,
-      startedAt: true,
-      endedAt: true,
-      durationSec: true,
-      totalDistanceM: true,
-      totalEnergyKcal: true,
-      avgHeartRate: true,
-      maxHeartRate: true,
-      createdAt: true,
-    },
-  });
+  // Independent reads — the sourcePriorityJson lookup doesn't depend on the
+  // row read, so run them concurrently rather than paying two sequential
+  // round trips.
+  const [rows, userRow] = await Promise.all([
+    prisma.workout.findMany({
+      where,
+      orderBy: [{ startedAt: "desc" }, { id: "asc" }],
+      select: {
+        id: true,
+        source: true,
+        externalId: true,
+        sportType: true,
+        startedAt: true,
+        endedAt: true,
+        durationSec: true,
+        totalDistanceM: true,
+        totalEnergyKcal: true,
+        avgHeartRate: true,
+        maxHeartRate: true,
+        createdAt: true,
+      },
+    }),
+    prisma.user.findUnique({
+      where: { id: userId },
+      select: { sourcePriorityJson: true },
+    }),
+  ]);

-  const userRow = await prisma.user.findUnique({
-    where: { id: userId },
-    select: { sourcePriorityJson: true },
-  });
  const canonical = pickCanonicalWorkoutRows(
    rows,
    userRow?.sourcePriorityJson ?? null,
@@ -154,27 +193,37 @@ async function buildWorkoutsResponse(
    return a.id.localeCompare(b.id);
  });

-  const page = canonical.slice(offset, offset + limit).map((row) => ({
-    id: row.id,
-    sportType: row.sportType,
-    startedAt: row.startedAt,
-    endedAt: row.endedAt,
-    durationSec: row.durationSec,
-    distanceM: row.totalDistanceM,
-    activeEnergyKcal: row.totalEnergyKcal,
-    avgHr: row.avgHeartRate,
-    maxHr: row.maxHeartRate,
-    source: row.source,
-    externalId: row.externalId,
-  }));
+  return { canonical, rawCount: rows.length };
+}
+
+function sliceWorkoutsProjection(
+  projection: WorkoutsProjection,
+  limit: number,
+  offset: number,
+): WorkoutsResult {
+  const page = projection.canonical
+    .slice(offset, offset + limit)
+    .map((row) => ({
+      id: row.id,
+      sportType: row.sportType,
+      startedAt: row.startedAt,
+      endedAt: row.endedAt,
+      durationSec: row.durationSec,
+      distanceM: row.totalDistanceM,
+      activeEnergyKcal: row.totalEnergyKcal,
+      avgHr: row.avgHeartRate,
+      maxHr: row.maxHeartRate,
+      source: row.source,
+      externalId: row.externalId,
+    }));

  return {
    workouts: page,
    meta: {
-      total: canonical.length,
+      total: projection.canonical.length,
      limit,
      offset,
-      droppedDuplicates: rows.length - canonical.length,
+      droppedDuplicates: projection.rawCount - projection.canonical.length,
    },
  };
 }
@@ -207,15 +256,18 @@ export const GET = apiHandler(async (request: NextRequest) => {
  const until = searchParams.get("until");
  const sportType = searchParams.get("sportType");

-  const params: WorkoutsParams = { limit, offset, since, until, sportType };
-  const cacheKey = `${user.id}|${limit}|${offset}|${since ?? ""}|${until ?? ""}|${sportType ?? ""}`;
+  // Deliberately excludes `limit`/`offset` — every page of one filter
+  // combination shares this cached projection (see the docblock above).
+  const projectionKey = `${user.id}|${since ?? ""}|${until ?? ""}|${sportType ?? ""}`;

-  const result = await cached(
-    caches.workouts as ServerCache<WorkoutsResult>,
-    cacheKey,
-    () => buildWorkoutsResponse(user.id, params),
+  const projection = await cached(
+    caches.workouts as ServerCache<WorkoutsProjection>,
+    projectionKey,
+    () => buildWorkoutsProjection(user.id, { since, until, sportType }),
    annotate,
  );

+  const result = sliceWorkoutsProjection(projection, limit, offset);
+
  return apiSuccess(result);
 });
@@ -4,6 +4,7 @@ import { useEffect, useMemo, useState } from "react";
 import Link from "next/link";
 import { useRouter } from "next/navigation";
 import {
+  Loader2,
  MessagesSquare,
  Paperclip,
  Search,
@@ -18,13 +19,15 @@ import { PageHeader } from "@/components/ui/page-header";
 import { QueryErrorCard } from "@/components/ui/query-error-card";
 import { COACH_SCROLLBAR } from "@/components/insights/coach-panel/message-thread";
 import {
-  useCoachConversations,
+  useCoachConversationHistory,
  useDeleteCoachConversationWithUndo,
 } from "@/components/insights/coach-panel/use-coach";
 import type { CoachConversationDTO } from "@/lib/ai/coach/types";
 import { useCoachLaunch } from "@/lib/insights/coach-launch-context";
+import { useDebouncedValue } from "@/hooks/use-debounced-value";
 import { useFeatureFlags } from "@/hooks/use-feature-flags";
 import { useDisableCoach } from "@/hooks/use-disable-coach";
+import { useLoadMoreSentinel } from "@/hooks/use-load-more-sentinel";
 import { useTranslations } from "@/lib/i18n/context";
 import { formatRelativeTime } from "@/lib/i18n/relative-time";
 import { cn } from "@/lib/utils";
@@ -37,8 +40,14 @@ import { cn } from "@/lib/utils";
 * sibling of the Coach page itself: a search field at the top, then the
 * recent conversations grouped by recency (Today / Yesterday / This week
 * / Earlier). Selecting a row routes to `/coach?c=<id>` — the existing
- * deep-link open mechanism — so there is NO new API and NO new state;
- * the page is a thin read over `useCoachConversations()`.
+ * deep-link open mechanism. Reuses the SAME `/coach?c=<id>` open
+ * mechanism the rail uses.
+ *
+ * v1.30.2 (QoL H1) — the page is a thin read over
+ * `useCoachConversationHistory()` (cursor-paginated + server-searched, see
+ * the route's doc comment), which the drawer's `<HistoryRail>` also
+ * consumes — the two surfaces can no longer drift onto different
+ * pagination behaviour.
 *
 * Gating mirrors `/coach`: the operator master flag OR a per-user opt-out
 * hides the Coach entirely and redirects back to `/insights` rather than
@@ -101,21 +110,42 @@ function groupByRecency(conversations: CoachConversationDTO[]): RecencyGroup[] {
 function CoachConversationsBody() {
  const { t } = useTranslations();
  const router = useRouter();
-  const { conversations, isLoading, isError, refetch } =
-    useCoachConversations();
+  const [filter, setFilter] = useState<string>("");
+  // v1.30.2 (QoL H1) — search now drives the server-side title query;
+  // debounce the keystroke draft instead of re-issuing a request per key.
+  const debouncedFilter = useDebouncedValue(filter, 200);
+  const {
+    conversations,
+    isLoading,
+    isError,
+    refetch,
+    hasNextPage,
+    isFetchingNextPage,
+    fetchNextPage,
+  } = useCoachConversationHistory({ search: debouncedFilter });
  const { pendingDeleteIds, requestDelete, undoDelete } =
    useDeleteCoachConversationWithUndo();
-  const [filter, setFilter] = useState<string>("");

-  // Mirror the rail's title substring filter.
-  const filtered = useMemo(() => {
-    const q = filter.trim().toLowerCase();
-    const visible = conversations.filter((c) => !pendingDeleteIds.has(c.id));
-    if (!q) return visible;
-    return visible.filter((c) => c.title.toLowerCase().includes(q));
-  }, [conversations, filter, pendingDeleteIds]);
+  // pendingDeleteIds is a client-only hide filter for the just-armed undo
+  // window — the server already resolved the rest of `visible` against the
+  // search term.
+  const visible = useMemo(
+    () => conversations.filter((c) => !pendingDeleteIds.has(c.id)),
+    [conversations, pendingDeleteIds],
+  );
+  const isSearching = debouncedFilter.trim().length > 0;

-  const groups = useMemo(() => groupByRecency(filtered), [filtered]);
+  const groups = useMemo(() => groupByRecency(visible), [visible]);
+
+  // Same callback-ref pattern as the rail: a `useState` setter (not a plain
+  // `useRef`) so the sentinel hook's `root` re-fires its effect once the
+  // real scroll-container node exists (null on the very first render).
+  const [listNode, setListNode] = useState<HTMLDivElement | null>(null);
+  const sentinelRef = useLoadMoreSentinel({
+    enabled: hasNextPage && !isFetchingNextPage,
+    onLoadMore: fetchNextPage,
+    root: listNode,
+  });

  function handleSelect(id: string) {
    // Reuse the existing `?c=<id>` open mechanism on the Coach page.
@@ -175,20 +205,21 @@ function CoachConversationsBody() {
      </header>

      <div
+        ref={setListNode}
        data-slot="coach-conversations-list"
        className={cn(
          "-mx-1 flex min-h-0 flex-1 flex-col gap-4 overflow-y-auto px-1",
          COACH_SCROLLBAR,
        )}
      >
-        {isLoading && conversations.length === 0 ? (
+        {isLoading && visible.length === 0 ? (
          <p
            data-slot="coach-conversations-loading"
            className="text-muted-foreground px-1 py-3 text-sm"
          >
            {t("common.loading")}
          </p>
-        ) : isError && conversations.length === 0 ? (
+        ) : isError && visible.length === 0 ? (
          // A load failure must not masquerade as "no conversations yet" —
          // that reads as a data-loss scare and offers no way back. Mirror
          // /coach/plans and surface a retry.
@@ -202,7 +233,9 @@ function CoachConversationsBody() {
              <MessagesSquare className="size-6" aria-hidden="true" />
            </span>
            <p className="text-muted-foreground max-w-xs text-sm leading-relaxed">
-              {t("insights.coach.historyEmpty")}
+              {isSearching
+                ? t("insights.coach.historySearchEmpty")
+                : t("insights.coach.historyEmpty")}
            </p>
          </div>
        ) : (
@@ -278,6 +311,26 @@ function CoachConversationsBody() {
            </section>
          ))
        )}
+        {/* v1.30.2 (QoL H1) — same IntersectionObserver sentinel as the
+            drawer rail: scrolling this div into view pulls the next cursor
+            page. Empty + aria-hidden; the status row below is the a11y
+            signal. */}
+        {visible.length > 0 && hasNextPage ? (
+          <div ref={sentinelRef} aria-hidden="true" className="h-px" />
+        ) : null}
+        {isFetchingNextPage ? (
+          <div
+            data-slot="coach-conversations-loading-more"
+            role="status"
+            className="text-muted-foreground flex items-center justify-center gap-2 py-3 text-xs"
+          >
+            <Loader2
+              className="size-3.5 animate-spin motion-reduce:animate-none"
+              aria-hidden="true"
+            />
+            {t("insights.coach.historyLoadingMore")}
+          </div>
+        ) : null}
      </div>
    </div>
  );
@@ -6,11 +6,13 @@ import { useQuery } from "@tanstack/react-query";

 import { CoachConversation } from "@/components/insights/coach-panel/coach-conversation";
 import type { CoachNudgeStatus } from "@/components/insights/layout-coach-fab";
+import type { CoachLaunchScope } from "@/lib/insights/coach-launch-context";
 import { useCoachLaunch } from "@/lib/insights/coach-launch-context";
 import { useFeatureFlags } from "@/hooks/use-feature-flags";
 import { useDisableCoach } from "@/hooks/use-disable-coach";
 import { queryKeys } from "@/lib/query-keys";
 import { apiGet } from "@/lib/api/api-fetch";
+import { coachScopeSourceSchema } from "@/lib/ai/coach/types";

 /**
 * v1.12.0 (Coach v2 #6) — full-page Coach conversation.
@@ -57,6 +59,23 @@ import { apiGet } from "@/lib/api/api-fetch";
 *     most-recent thread so the user lands on what the Coach said.
 * The search-param read sits in a Suspense child so the client-bailout
 * never de-opts the route.
+ *
+ * 2026-07-17 UX-flows audit — the route understood only `?c=` / `?doc=`, so
+ * every other cross-surface hand-off (a metric card's "ask about this", the
+ * Today check-in's "Adjust", a workout detail's coach button) dropped its
+ * context at the URL boundary and landed on a blank composer (F1-2 / F4-1 /
+ * F6-1). Two additive params close that gap, both seeding the SAME props
+ * `<CoachConversation>` already exposes for the drawer's suggested-prompt
+ * chips — no new plumbing inside the conversation surface itself:
+ *   - `?scope=<CoachScopeSource>` — narrows the snapshot the FIRST turn of a
+ *     fresh conversation reads (validated against the closed enum; an
+ *     unrecognised value is silently dropped rather than reaching the chat
+ *     route with a free-form string).
+ *   - `?ask=<text>` — seeds the composer prefill (mirrors `prefill` on the
+ *     in-app `askCoach()` launch call). The user still reviews/sends it —
+ *     this is a prefill, not an auto-send.
+ * Both are ignored once `?c=` or `?doc=` pin an existing/scoped thread — scope
+ * only ever applies to a fresh conversation's first turn.
 */
 function CoachPageBody() {
  const searchParams = useSearchParams();
@@ -71,13 +90,30 @@ function CoachPageBody() {
  const rawDoc = searchParams.get("doc");
  const seedDocumentId = deepLinkedId === null && rawDoc ? rawDoc : null;

+  // A fresh conversation only — an existing thread (`?c=`) or a doc-scoped
+  // chat (`?doc=`) keeps its own established scope.
+  const freshChat = deepLinkedId === null && seedDocumentId === null;
+  const rawScope = freshChat ? searchParams.get("scope") : null;
+  const scopeResult = rawScope
+    ? coachScopeSourceSchema.safeParse(rawScope)
+    : null;
+  const launchScope: CoachLaunchScope | null = scopeResult?.success
+    ? { metric: scopeResult.data }
+    : null;
+  const seedPrefill = freshChat ? searchParams.get("ask") : null;
+
  // C1 exception — an unread coach-initiated message opens the most-recent
  // conversation (which holds that proactive turn). Only consulted when the
  // entry did not pin a specific thread or ask for a fresh chat.
  // A `?doc=` open is an explicit fresh doc-scoped chat — never override it by
-  // resuming the most-recent thread on an unread nudge.
+  // resuming the most-recent thread on an unread nudge. A `?scope=`/`?ask=`
+  // hand-off is likewise an explicit fresh-chat request.
  const exceptionEligible =
-    deepLinkedId === null && rawC !== "new" && seedDocumentId === null;
+    deepLinkedId === null &&
+    rawC !== "new" &&
+    seedDocumentId === null &&
+    launchScope === null &&
+    !seedPrefill;
  const { data: nudge } = useQuery({
    queryKey: queryKeys.coachNudgeStatus(),
    queryFn: async (): Promise<CoachNudgeStatus> =>
@@ -95,6 +131,10 @@ function CoachPageBody() {
      // v1.28.51 — seed the document scope for a `?doc=<id>` open so the first
      // turn is created + sent through the hardened fenced document endpoint.
      initialDocumentId={seedDocumentId}
+      // 2026-07-17 UX-flows audit F1-2/F4-1/F6-1 — seed the scope/prefill a
+      // cross-surface hand-off carried in the URL.
+      launchScope={launchScope}
+      prefill={seedPrefill}
      // Default is the new-chat hero; only resume most-recent for the
      // unread coach-initiated exception.
      autoOpenMostRecent={hasUnreadCoachMessage}
@@ -3,7 +3,7 @@
 import { useQuery } from "@tanstack/react-query";
 import dynamic from "next/dynamic";
 import Link from "next/link";
-import { Smile } from "lucide-react";
+import { ArrowRight, Smile } from "lucide-react";

 import { useAuth } from "@/hooks/use-auth";
 import { queryKeys } from "@/lib/query-keys";
@@ -41,7 +41,7 @@ interface ComprehensiveMoodData {
 }

 export default function InsightsStimmungPage() {
-  const { isAuthenticated } = useAuth();
+  const { user, isAuthenticated } = useAuth();
  const { t } = useTranslations();
  const { compareBaseline } = useInsightsLayoutPrefs(isAuthenticated);

@@ -101,6 +101,32 @@ export default function InsightsStimmungPage() {
          The heatmap and the assessment are lifted out of `MoodInsightsSections`
          into their own regions so they land in this exact order; all three
          regions share one `moodInsights` query (TanStack dedups the fetch). */}
+
+      {/* 2026-07-17 UX/IA audit M9 — mood, the mental-wellbeing screeners, and
+          this insights page form one mental-health domain; this cross-links
+          into the other two so the trend a user reads here can lead
+          somewhere. */}
+      <div className="flex flex-wrap items-center gap-x-4 gap-y-1.5">
+        <Link
+          href="/mood"
+          data-slot="insights-mood-log-link"
+          className="text-muted-foreground hover:text-foreground focus-visible:ring-ring/50 inline-flex items-center gap-1.5 text-sm underline-offset-4 transition-colors hover:underline focus-visible:ring-[3px] focus-visible:outline-none"
+        >
+          {t("insights.mood.logLink")}
+          <ArrowRight className="size-3.5" aria-hidden="true" />
+        </Link>
+        {user?.modules?.mentalHealth === true ? (
+          <Link
+            href="/mental-wellbeing"
+            data-slot="insights-mood-mental-wellbeing-link"
+            className="text-muted-foreground hover:text-foreground focus-visible:ring-ring/50 inline-flex items-center gap-1.5 text-sm underline-offset-4 transition-colors hover:underline focus-visible:ring-[3px] focus-visible:outline-none"
+          >
+            {t("insights.mood.mentalWellbeingLink")}
+            <ArrowRight className="size-3.5" aria-hidden="true" />
+          </Link>
+        ) : null}
+      </div>
+
      <MoodInsightsSections region="heatmap" />

      {/* No `<MetricRangeControls>` here: mood is event-driven, not a
@@ -1,6 +1,7 @@
 "use client";

-import { Loader2 } from "lucide-react";
+import Link from "next/link";
+import { ArrowRight, Loader2 } from "lucide-react";

 import { useTranslations } from "@/lib/i18n/context";
 import { useModulePageGuard } from "@/hooks/use-module-page-guard";
@@ -46,6 +47,20 @@ export default function InsightsRecoveryPage() {
      explainerMetric="recoveryPage"
      coachLaunch
    >
+      {/* 2026-07-17 UX/IA audit M4 — reverse direction of the cross-link added
+          to the score-anatomy page: this composite wearable page shares the
+          "Recovery" nav label with the RECOVERY_SCORE anatomy view at
+          `/insights/scores/recovery`, so a user who arrived here looking for
+          the score ring / contributor breakdown gets a pointer instead of a
+          dead end. */}
+      <Link
+        href="/insights/scores/recovery"
+        data-slot="recovery-anatomy-cross-link"
+        className="text-muted-foreground hover:text-foreground focus-visible:ring-ring/50 -mt-2 inline-flex items-center gap-1.5 text-sm transition-colors focus-visible:ring-[3px] focus-visible:outline-none"
+      >
+        {t("insights.recovery.scoreCrossLink")}
+        <ArrowRight className="size-3.5" aria-hidden="true" />
+      </Link>
      <RecoverySection />
    </SubPageShell>
  );
@@ -1,6 +1,8 @@
 "use client";

 import { useParams } from "next/navigation";
+import Link from "next/link";
+import { ArrowRight } from "lucide-react";

 import { useTranslations } from "@/lib/i18n/context";
 import { BackLink } from "@/components/ui/back-link";
@@ -94,6 +96,22 @@ export default function CompositeScorePage() {
      }
      coachLaunch
    >
+      {/* 2026-07-17 UX/IA audit M4 — this anatomy view and the `/insights/recovery`
+          composite-wearable page both used to render as plain "Recovery", a
+          same-name/different-content trap: the score ring here promised one
+          thing, the nav pill led somewhere else with no pointer back. The
+          label above is now "Recovery score" (matches `measurements.typeRecoveryScore`);
+          this line closes the loop with an explicit cross-link. */}
+      {metric === "RECOVERY_SCORE" ? (
+        <Link
+          href="/insights/recovery"
+          data-slot="recovery-score-cross-link"
+          className="text-muted-foreground hover:text-foreground focus-visible:ring-ring/50 -mt-2 inline-flex items-center gap-1.5 text-sm transition-colors focus-visible:ring-[3px] focus-visible:outline-none"
+        >
+          {t("insights.derived.scores.recoveryCrossLink")}
+          <ArrowRight className="size-3.5" aria-hidden="true" />
+        </Link>
+      ) : null}
      <CompositeScoreAnatomy metric={metric} />
    </SubPageShell>
  );
@@ -78,7 +78,11 @@ export default function InsightsWorkoutDetailPage({
          <WorkoutDetailStats workout={data} />
          <WorkoutDetailRoute workout={data} />
          <WorkoutDetailHRChart workout={data} />
-          <CoachLaunchButton />
+          {/* 2026-07-17 UX-flows audit F6-1 — the button used to open an
+              unscoped chat; the user had to restate which session they
+              meant. `workouts` narrows the snapshot the first turn reads
+              (pairs with the coach-scope param added for F4-1). */}
+          <CoachLaunchButton scope={{ metric: "workouts" }} />
        </>
      )}
    </SubPageShell>
@@ -13,7 +13,7 @@ import { PageHeader } from "@/components/ui/page-header";
 import { PageAuthGate } from "@/components/ui/page-auth-gate";
 import { PullToRefreshIndicator } from "@/components/ui/pull-to-refresh-indicator";
 import { ResponsiveSheet } from "@/components/ui/responsive-sheet";
-import { Plus, Wrench } from "lucide-react";
+import { ArrowRight, Plus, Wrench } from "lucide-react";
 import { useRouter } from "next/navigation";
 import { useEffect } from "react";
 import { useTranslations } from "@/lib/i18n/context";
@@ -120,6 +120,22 @@ export default function MoodPage() {
      </ResponsiveSheet>

      <MoodList onAddFirst={() => setDialogOpen(true)} />
+
+      {/* 2026-07-17 UX/IA audit M9 — mood tracking, the mental-wellbeing
+          screeners, and the mood insights page form one mental-health
+          domain but used to be three unconnected islands. A quiet pointer
+          here (module-gated — no nav change) rather than a merge: capture
+          vs. screener stay two defensible surfaces, just cross-linked. */}
+      {user?.modules?.mentalHealth === true ? (
+        <Link
+          href="/mental-wellbeing"
+          data-slot="mood-mental-wellbeing-link"
+          className="text-muted-foreground hover:text-foreground focus-visible:ring-ring/50 inline-flex items-center gap-1.5 text-sm underline-offset-4 transition-colors hover:underline focus-visible:ring-[3px] focus-visible:outline-none"
+        >
+          {t("mood.mentalWellbeingLink")}
+          <ArrowRight className="size-3.5" aria-hidden="true" />
+        </Link>
+      ) : null}
    </div>
  );
 }
@@ -18,8 +18,10 @@
 */
 import { useMutation, useQuery, useQueryClient } from "@tanstack/react-query";
 import {
+  ArrowRight,
  Check,
  Download,
+  FlaskConical,
  Pencil,
  Plus,
  Share2,
@@ -32,6 +34,7 @@ import { useEffect, useRef, useState } from "react";
 import { toast } from "sonner";

 import { useIllnessEpisodes } from "@/components/illness/use-illness";
+import { useAuth } from "@/hooks/use-auth";
 import { Button } from "@/components/ui/button";
 import { DateField } from "@/components/ui/date-field";
 import { Input } from "@/components/ui/input";
@@ -230,6 +233,21 @@ export function DocumentDetailSheet({
  });
  const doc = detail.data;

+  // 2026-07-17 UX/IA audit M5 / F5-5 — the document → lab-values direction
+  // had no UI link at all: an auto-staged (PENDING) or OCR-confirmed
+  // (APPROVED, see `linkOcrLabsToVaultDocument`) OBSERVATION fact already
+  // ties this document to specific lab rows, but nothing on the sheet said
+  // so. There is no per-value route yet (`/labs` has no `?analyte=`
+  // deep-link — see the labs page's own "no per-value relation" comment),
+  // so this links to the labs list rather than a single biomarker; still a
+  // real jump where there was none before.
+  const labFactCount =
+    doc?.facts.filter(
+      (fact) => fact.factType === "OBSERVATION" && fact.status !== "REJECTED",
+    ).length ?? 0;
+  const { user } = useAuth();
+  const labsModuleEnabled = user?.modules?.labs !== false;
+
  const episodes = useIllnessEpisodes(true);

  // AI layer — capability drives the vision/text transport; the mutations run
@@ -792,6 +810,18 @@ export function DocumentDetailSheet({
                </div>
              </div>

+              {labsModuleEnabled && labFactCount > 0 ? (
+                <Link
+                  href="/labs"
+                  data-slot="document-detail-lab-values-link"
+                  className="text-muted-foreground hover:text-foreground focus-visible:ring-ring/50 inline-flex items-center gap-1.5 self-start rounded-md text-sm transition-colors focus-visible:ring-[3px] focus-visible:outline-none"
+                >
+                  <FlaskConical className="size-4" aria-hidden="true" />
+                  {t("documents.detail.labValuesLink", { count: labFactCount })}
+                  <ArrowRight className="size-3.5" aria-hidden="true" />
+                </Link>
+              ) : null}
+
              <p className="text-muted-foreground text-xs">
                {t("documents.detail.uploadedMeta", {
                  date: format.date(doc.createdAt),
@@ -5,6 +5,7 @@ import { Sparkles } from "lucide-react";
 import { Button } from "@/components/ui/button";
 import { cn } from "@/lib/utils";
 import { useTranslations } from "@/lib/i18n/context";
+import type { CoachLaunchScope } from "@/lib/insights/coach-launch-context";
 import { useCoachLaunch } from "@/lib/insights/coach-launch-context";
 import { useFeatureFlags } from "@/hooks/use-feature-flags";
 import { useDisableCoach } from "@/hooks/use-disable-coach";
@@ -31,6 +32,13 @@ export interface CoachLaunchButtonProps {
  label?: string;
  /** Optional prefill seed for the next Coach turn. */
  prefill?: string;
+  /**
+   * 2026-07-17 UX-flows audit F6-1 — optional scope so the conversation this
+   * button opens narrows to the relevant source(s) instead of starting from
+   * a blank, unscoped chat. Passed straight through to `askCoach`'s second
+   * argument (mirrors every other ambient-scope call site).
+   */
+  scope?: CoachLaunchScope;
  /** Optional className passthrough for inline overrides. */
  className?: string;
  /**
@@ -49,6 +57,7 @@ export interface CoachLaunchButtonProps {
 export function CoachLaunchButton({
  label,
  prefill,
+  scope,
  className,
  variant = "inline",
 }: CoachLaunchButtonProps) {
@@ -83,7 +92,7 @@ export function CoachLaunchButton({
        data-slot="coach-launch-icon"
        aria-label={accessibleLabel}
        title={accessibleLabel}
-        onClick={() => launch.askCoach(prefill ?? null)}
+        onClick={() => launch.askCoach(prefill ?? null, scope)}
        className={cn("text-muted-foreground hover:text-foreground", className)}
      >
        <Sparkles className="size-4" aria-hidden="true" />
@@ -101,7 +110,7 @@ export function CoachLaunchButton({
      variant="outline"
      size="sm"
      data-slot="coach-launch-inline"
-      onClick={() => launch.askCoach(prefill ?? null)}
+      onClick={() => launch.askCoach(prefill ?? null, scope)}
      className={cn("inline-flex h-10 gap-2 self-end", className)}
    >
      <Sparkles className="size-4" aria-hidden="true" />
@@ -3,23 +3,46 @@ import { renderToStaticMarkup } from "react-dom/server";
 import { QueryClient, QueryClientProvider } from "@tanstack/react-query";

 import { I18nProvider } from "@/lib/i18n/context";
+import { queryKeys } from "@/lib/query-keys";
 import { HistoryRail } from "../history-rail";
 import type { CoachConversationsPage } from "@/lib/ai/coach/types";

 /**
 * v1.4.20 phase B2b — history-rail render snapshot.
+ * v1.30.2 (QoL H1) — the rail now reads `useCoachConversationHistory`
+ * (`useInfiniteQuery`), so the seeded cache shape moved from a single
+ * `CoachConversationsPage` under the flat `coachConversations()` key to the
+ * `{ pages, pageParams }` infinite-query shape under
+ * `queryKeys.coachConversationHistory(search)`.
 *
 * The rail mounts a TanStack Query subscription so we have to wrap it
- * in a `<QueryClientProvider>`. We seed the cache with a synthetic
- * conversations page so the rail renders the populated state in one
- * pass — server-side rendering doesn't drive a real network call.
+ * in a `<QueryClientProvider>`. We seed the cache with synthetic page(s)
+ * so the rail renders the populated state in one pass — server-side
+ * rendering doesn't drive a real network call.
 */

-function makeClientWithConversations(page: CoachConversationsPage) {
+function makeClientWithConversations(
+  page: CoachConversationsPage,
+  search = "",
+) {
  const client = new QueryClient({
    defaultOptions: { queries: { retry: false } },
  });
-  client.setQueryData(["coachConversations"], page);
+  client.setQueryData(queryKeys.coachConversationHistory(search), {
+    pages: [page],
+    pageParams: [null],
+  });
+  return client;
+}
+
+function makeClientWithPages(pages: CoachConversationsPage[], search = "") {
+  const client = new QueryClient({
+    defaultOptions: { queries: { retry: false } },
+  });
+  client.setQueryData(queryKeys.coachConversationHistory(search), {
+    pages,
+    pageParams: pages.map((_, i) => (i === 0 ? null : `cursor-${i}`)),
+  });
  return client;
 }

@@ -152,4 +175,55 @@ describe("<HistoryRail>", () => {
    const buttons = html.match(/data-slot="coach-history-delete"/g) ?? [];
    expect(buttons.length).toBe(2);
  });
+
+  // v1.30.2 (QoL H1) — infinite history: reachability beyond the first page.
+  it("renders every conversation across multiple cached pages (not just page one)", () => {
+    const client = makeClientWithPages([
+      samplePage,
+      {
+        conversations: [
+          {
+            id: "c3",
+            title: "A much older thread",
+            createdAt: "2026-01-01T08:00:00.000Z",
+            updatedAt: "2026-01-01T08:01:00.000Z",
+            messageCount: 1,
+            fenced: false,
+          },
+        ],
+        nextCursor: null,
+      },
+    ]);
+    const html = render(
+      <HistoryRail activeId={null} onSelect={() => {}} />,
+      client,
+    );
+    const rows = html.match(/data-slot="coach-history-item"/g) ?? [];
+    expect(rows.length).toBe(3);
+    expect(html).toContain("A much older thread");
+  });
+
+  it("renders the load-more sentinel when a cached page still carries a nextCursor", () => {
+    const client = makeClientWithConversations({
+      ...samplePage,
+      nextCursor: "c2",
+    });
+    const html = render(
+      <HistoryRail activeId={null} onSelect={() => {}} />,
+      client,
+    );
+    // The sentinel div is aria-hidden — IntersectionObserver never fires
+    // during a static SSR render, but the element itself must be present
+    // for the rail to ever pull the next page in a real browser.
+    expect(html).toMatch(/aria-hidden="true"[^>]*class="h-px"/);
+  });
+
+  it("does not render the load-more sentinel once the caller has reached the end", () => {
+    const client = makeClientWithConversations(samplePage); // nextCursor: null
+    const html = render(
+      <HistoryRail activeId={null} onSelect={() => {}} />,
+      client,
+    );
+    expect(html).not.toMatch(/class="h-px"/);
+  });
 });
@@ -1,7 +1,14 @@
 "use client";

 import { useMemo, useState } from "react";
-import { MessagesSquare, Paperclip, Search, Trash2 } from "lucide-react";
+import {
+  Loader2,
+  MessagesSquare,
+  Paperclip,
+  RotateCw,
+  Search,
+  Trash2,
+} from "lucide-react";
 import { toast } from "sonner";

 import { Button } from "@/components/ui/button";
@@ -9,10 +16,12 @@ import { Input } from "@/components/ui/input";
 import { cn } from "@/lib/utils";
 import { useTranslations } from "@/lib/i18n/context";
 import { formatRelativeTime } from "@/lib/i18n/relative-time";
+import { useDebouncedValue } from "@/hooks/use-debounced-value";
+import { useLoadMoreSentinel } from "@/hooks/use-load-more-sentinel";

 import { COACH_SCROLLBAR } from "./message-thread";
 import {
-  useCoachConversations,
+  useCoachConversationHistory,
  useDeleteCoachConversationWithUndo,
 } from "./use-coach";

@@ -20,9 +29,10 @@ import {
 * v1.4.20 phase B2b — conversation history rail.
 *
 * Lives in the left column of the Coach drawer on `lg+`. Renders the
- * paginated list returned by `GET /api/insights/chat`, lets the user
- * filter by title (substring match), select one to resume, and delete
- * with a confirm-then-go flow.
+ * caller's FULL conversation history (v1.30.2 — cursor-paginated via
+ * `useInfiniteQuery`, loaded incrementally as the list scrolls near its
+ * end), lets the user search by title (server-side, debounced 200 ms),
+ * select one to resume, and delete with an undo-able confirm.
 *
 * Selection / activeId is a controlled prop so the drawer owns the
 * authoritative state across the message thread + composer.
@@ -55,17 +65,41 @@ export function HistoryRail({
  hideHeading = false,
 }: HistoryRailProps) {
  const { t } = useTranslations();
-  const { conversations, isLoading } = useCoachConversations();
+  const [filter, setFilter] = useState<string>("");
+  // v1.30.2 (QoL H1) — the search box now drives a SERVER-side query
+  // (title-only substring, see the route doc comment), so debounce the
+  // keystroke draft rather than re-issuing a request per character.
+  const debouncedFilter = useDebouncedValue(filter, 200);
+  const {
+    conversations,
+    isLoading,
+    isError,
+    refetch,
+    hasNextPage,
+    isFetchingNextPage,
+    fetchNextPage,
+  } = useCoachConversationHistory({ search: debouncedFilter });
  const { pendingDeleteIds, requestDelete, undoDelete } =
    useDeleteCoachConversationWithUndo();
-  const [filter, setFilter] = useState<string>("");

-  const filtered = useMemo(() => {
-    const q = filter.trim().toLowerCase();
-    const visible = conversations.filter((c) => !pendingDeleteIds.has(c.id));
-    if (!q) return visible;
-    return visible.filter((c) => c.title.toLowerCase().includes(q));
-  }, [conversations, filter, pendingDeleteIds]);
+  // pendingDeleteIds is a client-only hide filter for the just-armed undo
+  // window — the server-side search already resolved the rest of `visible`.
+  const visible = useMemo(
+    () => conversations.filter((c) => !pendingDeleteIds.has(c.id)),
+    [conversations, pendingDeleteIds],
+  );
+  const isSearching = debouncedFilter.trim().length > 0;
+
+  // A `useState`-backed callback ref (not a plain `useRef`) so the
+  // container's assignment itself triggers a re-render — the sentinel
+  // hook's effect depends on `root` and must re-run once the real DOM
+  // node exists (it's null during the very first render pass).
+  const [listNode, setListNode] = useState<HTMLDivElement | null>(null);
+  const sentinelRef = useLoadMoreSentinel({
+    enabled: hasNextPage && !isFetchingNextPage,
+    onLoadMore: fetchNextPage,
+    root: listNode,
+  });

  // v1.30.1 M5 — a single tap hides the row and schedules the real
  // delete; the toast's Undo action cancels it within the grace
@@ -126,6 +160,7 @@ export function HistoryRail({
        />
      </div>
      <div
+        ref={setListNode}
        data-slot="coach-history-list"
        className={cn(
          "-mx-1 flex min-h-0 flex-1 flex-col gap-0.5 overflow-y-auto px-1",
@@ -133,22 +168,37 @@ export function HistoryRail({
          COACH_SCROLLBAR,
        )}
      >
-        {isLoading && conversations.length === 0 ? (
+        {isLoading && visible.length === 0 ? (
          <p
            data-slot="coach-history-loading"
            className="text-muted-foreground px-2 py-3 text-sm"
          >
            {t("common.loading")}
          </p>
-        ) : filtered.length === 0 ? (
+        ) : isError && visible.length === 0 ? (
+          // A load failure must not masquerade as "no conversations yet" —
+          // offer a retry instead of the empty copy.
+          <div
+            data-slot="coach-history-error"
+            className="text-muted-foreground flex flex-col items-start gap-2 px-2 py-3 text-sm"
+          >
+            <p>{t("common.loadFailed")}</p>
+            <Button variant="outline" size="sm" onClick={() => refetch()}>
+              <RotateCw className="size-3.5" aria-hidden="true" />
+              {t("common.retry")}
+            </Button>
+          </div>
+        ) : visible.length === 0 ? (
          <p
            data-slot="coach-history-empty"
            className="text-muted-foreground px-2 py-3 text-sm leading-relaxed"
          >
-            {t("insights.coach.historyEmpty")}
+            {isSearching
+              ? t("insights.coach.historySearchEmpty")
+              : t("insights.coach.historyEmpty")}
          </p>
        ) : (
-          filtered.map((c) => {
+          visible.map((c) => {
            const isActive = c.id === activeId;
            return (
              <div
@@ -214,6 +264,26 @@ export function HistoryRail({
            );
          })
        )}
+        {/* v1.30.2 (QoL H1) — IntersectionObserver sentinel: scrolling this
+            div into the rail's own scrollport pulls the next cursor page.
+            Empty + aria-hidden — the loading row right below it is the
+            a11y-visible signal. */}
+        {visible.length > 0 && hasNextPage ? (
+          <div ref={sentinelRef} aria-hidden="true" className="h-px" />
+        ) : null}
+        {isFetchingNextPage ? (
+          <div
+            data-slot="coach-history-loading-more"
+            role="status"
+            className="text-muted-foreground flex items-center justify-center gap-2 px-2 py-3 text-xs"
+          >
+            <Loader2
+              className="size-3.5 animate-spin motion-reduce:animate-none"
+              aria-hidden="true"
+            />
+            {t("insights.coach.historyLoadingMore")}
+          </div>
+        ) : null}
      </div>
    </div>
  );
@@ -1,7 +1,12 @@
 "use client";

-import { useCallback, useEffect, useRef, useState } from "react";
-import { useMutation, useQuery, useQueryClient } from "@tanstack/react-query";
+import { useCallback, useEffect, useMemo, useRef, useState } from "react";
+import {
+  useInfiniteQuery,
+  useMutation,
+  useQuery,
+  useQueryClient,
+} from "@tanstack/react-query";

 import type {
  CoachConversationAttachmentDTO,
@@ -41,10 +46,13 @@ const QUERY_KEYS = {
 };

 /**
- * List of conversations for the rail. Cursor pagination is exposed via
- * `loadMore` — the rail can call it when the user scrolls past the
- * end. v1.4.20 lands with default-page-only; cursors are pre-wired so
- * v1.4.21 / v1.5 can add infinite scroll without reshaping the hook.
+ * The single (first) page of conversations. Used ONLY for the
+ * auto-open-most-recent-thread behaviour on mount
+ * (`coach-conversation.tsx`), which never needs more than the rail's
+ * server-authoritative head. Every surface that lets the user BROWSE or
+ * SEARCH history (the drawer rail, the standalone conversations page) reads
+ * `useCoachConversationHistory` below instead — it walks the full cursor
+ * chain via `useInfiniteQuery` rather than stopping at page one.
 */
 export function useCoachConversations(enabled = true) {
  const query = useQuery({
@@ -67,6 +75,76 @@ export function useCoachConversations(enabled = true) {
  };
 }

+const HISTORY_PAGE_LIMIT = 20;
+
+export interface UseCoachConversationHistoryOptions {
+  /**
+   * Server-side title search (see `GET /api/insights/chat?q=`). The caller
+   * debounces the raw input (`useDebouncedValue`) before passing it here —
+   * every distinct value keys its own cache page chain, so a fast typist
+   * never sees a stale search's tail page bleed into a fresh one.
+   */
+  search?: string;
+  enabled?: boolean;
+}
+
+/**
+ * v1.30.2 (QoL H1) — the full, paginated + server-searched conversation
+ * history. Replaces the old "first page only, client-side substring
+ * filter" behaviour: `useInfiniteQuery` walks the cursor chain
+ * (`nextCursor`) so every conversation the caller has ever had is
+ * reachable via `fetchNextPage`, and `search` narrows the SERVER'S query
+ * (title-only — see the route doc comment) rather than filtering an
+ * already-truncated client array.
+ *
+ * Shared by the drawer's `<HistoryRail>` and the standalone
+ * `/coach/conversations` page so the two surfaces can never drift onto
+ * different pagination behaviour again.
+ */
+export function useCoachConversationHistory(
+  options: UseCoachConversationHistoryOptions = {},
+) {
+  const { search = "", enabled = true } = options;
+  const trimmedSearch = search.trim();
+
+  const query = useInfiniteQuery({
+    queryKey: queryKeys.coachConversationHistory(trimmedSearch),
+    queryFn: async ({
+      pageParam,
+    }: {
+      pageParam: string | null;
+    }): Promise<CoachConversationsPage> => {
+      const params = new URLSearchParams();
+      if (pageParam) params.set("cursor", pageParam);
+      params.set("limit", String(HISTORY_PAGE_LIMIT));
+      if (trimmedSearch) params.set("q", trimmedSearch);
+      return apiGet<CoachConversationsPage>(
+        `/api/insights/chat?${params.toString()}`,
+        { headers: { Accept: "application/json" } },
+      );
+    },
+    initialPageParam: null as string | null,
+    getNextPageParam: (lastPage) => lastPage.nextCursor,
+    enabled,
+    staleTime: 30 * 1000,
+  });
+
+  const conversations = useMemo(
+    () => query.data?.pages.flatMap((page) => page.conversations) ?? [],
+    [query.data],
+  );
+
+  return {
+    conversations,
+    isLoading: query.isLoading,
+    isError: query.isError,
+    hasNextPage: query.hasNextPage,
+    isFetchingNextPage: query.isFetchingNextPage,
+    fetchNextPage: query.fetchNextPage,
+    refetch: query.refetch,
+  };
+}
+
 /**
 * Fetch one conversation with every message decrypted server-side.
 * `id === null` disables the query so the drawer can mount without a
@@ -28,7 +28,9 @@
 * the page someone arrives at to take stock.
 */
 import { useMemo, useState } from "react";
+import Link from "next/link";
 import { useMutation, useQuery, useQueryClient } from "@tanstack/react-query";
+import { ArrowRight } from "lucide-react";

 import { useTranslations } from "@/lib/i18n/context";
 import { apiGet, apiPost } from "@/lib/api/api-fetch";
@@ -143,6 +145,19 @@ export function MentalWellbeing() {
            </ul>
          </section>

+          {/* 2026-07-17 UX/IA audit M9 — mood tracking, this screener surface,
+              and the mood insights page form one mental-health domain but used
+              to be three unconnected islands. A quiet pointer to mood history,
+              mirroring the one added on `/mood` pointing back here. */}
+          <Link
+            href="/mood"
+            data-slot="mental-wellbeing-mood-link"
+            className="text-muted-foreground hover:text-foreground focus-visible:ring-ring/50 -mt-2 inline-flex items-center gap-1.5 self-start text-sm underline-offset-4 transition-colors hover:underline focus-visible:ring-[3px] focus-visible:outline-none"
+          >
+            {t("mentalHealth.moodLink")}
+            <ArrowRight className="size-3.5" aria-hidden="true" />
+          </Link>
+
          {/* Per-instrument detail, opened from a card body: last result,
              trend chart, dated history, Start — the Verlauf lives here,
              not on the landing. */}
@@ -15,7 +15,11 @@ const integrationStatusState = vi.hoisted(() => ({
  data: undefined as
    | {
        threshold: number;
-        integrations: Array<{ integration: string; configured?: boolean }>;
+        integrations: Array<{
+          integration: string;
+          configured?: boolean;
+          connected?: boolean;
+        }>;
      }
    | undefined,
 }));
@@ -88,4 +92,23 @@ describe("<SourceCardGrid> Withings card", () => {
    expect(html).not.toContain('data-testid="source-card-withings-setup"');
    expect(html).toContain("Connect Withings");
  });
+
+  // 2026-07-17 UX-onboarding audit M6 — the connect CTA opens OAuth in a
+  // new tab; the callback flips the connection in THAT tab while this
+  // wizard tab used to render the same unchanged button forever. Once
+  // `useIntegrationStatuses` (refetches on window focus) reports
+  // `connected: true`, the card must swap to a badge instead of a dead
+  // "Connect Withings" button that would just reopen a redundant flow.
+  it("shows a connected badge instead of the connect button once Withings is linked", () => {
+    integrationStatusState.data = {
+      threshold: 3,
+      integrations: [
+        { integration: "withings", configured: true, connected: true },
+      ],
+    };
+    const html = render();
+    expect(html).toContain('data-testid="source-card-withings-connected"');
+    expect(html).not.toContain('href="/api/withings/connect"');
+    expect(html).not.toContain("Connect Withings");
+  });
 });
@@ -3,7 +3,14 @@
 import Link from "next/link";
 import { useEffect, useState } from "react";
 import { useQuery } from "@tanstack/react-query";
-import { CheckCircle2, FileUp, PlusCircle, Plug, Sparkles } from "lucide-react";
+import {
+  CheckCircle2,
+  FileUp,
+  PlusCircle,
+  Plug,
+  Settings2,
+  Sparkles,
+} from "lucide-react";

 import { Button } from "@/components/ui/button";
 import { MedicalDisclaimer } from "@/components/common/medical-disclaimer";
@@ -220,13 +227,29 @@ export function DoneScreen() {
        </Button>
      </div>

-      <Link
-        href="/settings/export"
-        className="text-muted-foreground hover:text-foreground inline-flex items-center gap-1.5 text-sm underline-offset-4 hover:underline"
-      >
-        <FileUp className="size-3.5" />
-        {t("onboarding.done.importCta")}
-      </Link>
+      {/* 2026-07-17 UX-onboarding audit M5 — the module system (default-on
+          set + opt-ins like nutrients / mental health / documents) is
+          otherwise undiscoverable during first-run: the wizard never
+          mentions it, and the only pointer used to be one tour stop.
+          A quiet link here, beside the existing import link, gives every
+          new account one honest pointer to Settings → Modules before they
+          leave the wizard. */}
+      <div className="flex flex-wrap items-center justify-center gap-x-4 gap-y-1.5">
+        <Link
+          href="/settings/export"
+          className="text-muted-foreground hover:text-foreground inline-flex items-center gap-1.5 text-sm underline-offset-4 hover:underline"
+        >
+          <FileUp className="size-3.5" />
+          {t("onboarding.done.importCta")}
+        </Link>
+        <Link
+          href="/settings/modules"
+          className="text-muted-foreground hover:text-foreground inline-flex items-center gap-1.5 text-sm underline-offset-4 hover:underline"
+        >
+          <Settings2 className="size-3.5" />
+          {t("onboarding.done.modulesCta")}
+        </Link>
+      </div>
    </section>
  );
 }
@@ -249,13 +249,19 @@ export function GettingStartedChecklist() {
    enabled: checklistRelevant,
  });

-  // v1.5 perf audit: skip these two fetches once the user is past
-  // onboarding. The checklist hides itself anyway via shouldShowChecklist
-  // when onboardingCompletedAt != null AND measurementCount >= 5;
+  // v1.5 perf audit: skip these two fetches once the card can't render at
+  // all (`checklistRelevant`, same gate the meds/AI queries above use).
  // withings/status and notifications/preferences are unique to this
  // component and were burning ~950 ms of network on every dashboard load
  // for established users. See docs/audit/v15-performance.md.
-  const onboardingPending = !!user && user.onboardingCompletedAt == null;
+  //
+  // 2026-07-17 UX audit M1 — this used to read a separate `onboardingPending`
+  // (`onboardingCompletedAt == null`) instead of `checklistRelevant`, so a
+  // user who finished the wizard but still had under 5 measurements kept
+  // seeing the card (`shouldShowChecklist` allows that) while these two
+  // queries stayed disabled forever: "Connect a data source" and
+  // "Configure notifications" could never flip done even after the user
+  // actually connected Withings or a push channel post-wizard.

  // v1.17.0 — any connected data source satisfies the step, so we read
  // the consolidated integrations envelope (Withings / WHOOP / Fitbit /
@@ -265,7 +271,7 @@ export function GettingStartedChecklist() {
    queryFn: async () => {
      return apiGet("/api/integrations/status");
    },
-    enabled: onboardingPending,
+    enabled: checklistRelevant,
  });

  const { data: notificationsData } = useQuery<NotificationsPreferences>({
@@ -273,7 +279,7 @@ export function GettingStartedChecklist() {
    queryFn: async () => {
      return apiGet("/api/notifications/preferences");
    },
-    enabled: onboardingPending,
+    enabled: checklistRelevant,
  });

  const medicationCount = medsData?.length ?? 0;
@@ -6,6 +6,7 @@ import { useRouter } from "next/navigation";
 import { useQueryClient } from "@tanstack/react-query";
 import { toast } from "sonner";
 import {
+  CheckCircle2,
  ClipboardCheck,
  FileUp,
  Plug,
@@ -119,6 +120,16 @@ export function SourceCardGrid() {
  const { data: integrationStatus } = useIntegrationStatuses(isAuthenticated);
  const withingsConfigured =
    pickStatus(integrationStatus, "withings")?.configured ?? false;
+  // 2026-07-17 UX-onboarding audit M6 — the connect CTA opens the OAuth
+  // handshake in a new tab; the callback lands THAT tab on
+  // Settings → Integrations while this wizard tab never learned the
+  // connection succeeded. `useIntegrationStatuses` already refetches on
+  // window focus, so returning to this tab after the OAuth round trip
+  // picks up the fresh `connected` flag for free — the only missing piece
+  // was rendering it as a badge instead of leaving the button looking
+  // untouched.
+  const withingsConnected =
+    pickStatus(integrationStatus, "withings")?.connected ?? false;

  async function advance() {
    if (advancing) return;
@@ -164,6 +175,8 @@ export function SourceCardGrid() {
            }
            withingsCtaLabel={t("onboarding.source.withings.cta")}
            withingsConfigured={withingsConfigured}
+            withingsConnected={withingsConnected}
+            withingsConnectedLabel={t("settings.integrationPill.connected")}
            withingsSetupCtaLabel={t("onboarding.source.withings.setupCta")}
            withingsSetupHint={t("onboarding.source.withings.setupHint")}
          />
@@ -260,6 +273,13 @@ interface SourceCardItemProps {
   *  credentials saved — i.e. whether `/api/withings/connect` can
   *  actually succeed right now. */
  withingsConfigured: boolean;
+  /** Whether Withings is already connected (OAuth completed in the tab the
+   *  connect CTA opened). Replaces the connect button with a badge so
+   *  returning to this wizard tab shows the outcome instead of an
+   *  unchanged card. */
+  withingsConnected: boolean;
+  /** Localised "Connected" label for the badge above. */
+  withingsConnectedLabel: string;
  /** CTA label shown instead of "Connect Withings" when unconfigured. */
  withingsSetupCtaLabel: string;
  /** One-sentence explanation of the BYO credential prerequisite. */
@@ -275,6 +295,8 @@ function SourceCardItem({
  recommendedLabel,
  withingsCtaLabel,
  withingsConfigured,
+  withingsConnected,
+  withingsConnectedLabel,
  withingsSetupCtaLabel,
  withingsSetupHint,
 }: SourceCardItemProps) {
@@ -325,7 +347,15 @@ function SourceCardItem({
      >
        {headerRow}
        {body}
-        {withingsConfigured ? (
+        {withingsConnected ? (
+          <div
+            className="text-primary flex items-center gap-1.5 pt-1 text-sm font-medium"
+            data-testid="source-card-withings-connected"
+          >
+            <CheckCircle2 className="size-3.5" aria-hidden="true" />
+            {withingsConnectedLabel}
+          </div>
+        ) : withingsConfigured ? (
          <div className="flex items-center gap-2 pt-1">
            <Button asChild variant="outline" size="sm">
              <a
@@ -19,6 +19,13 @@
 *
 * `parseOAuthOutcome` / `oauthReasonKey` are re-exported from
 * `connections-panel.tsx` so existing unit-test imports keep resolving.
+ *
+ * 2026-07-17 UX/IA audit M7 — `<NutrientIntakeCard>` (v1.28) moved here from
+ * Settings → Sources. It is a read-only sync inventory (what the Apple
+ * Health nutrients opt-in has stored), not a which-connection-wins ranking —
+ * "Source priority" describes the latter only. Matches the standing
+ * providers→Integrations rule; the card still renders nothing while the
+ * `nutrients` module is off.
 */

 import Link from "next/link";
@@ -28,6 +35,7 @@ import { useTranslations } from "@/lib/i18n/context";
 import { SettingsCardHeader } from "@/components/settings/_card-header";
 import { ConnectionsPanel } from "@/components/settings/integrations/connections-panel";
 import { NotificationChannelsPanel } from "@/components/settings/integrations/notification-channels-panel";
+import { NutrientIntakeCard } from "@/components/settings/nutrient-intake-card";
 import { Button } from "@/components/ui/button";

 export {
@@ -55,6 +63,15 @@ export function IntegrationsSection() {
          )}
        />
        <ConnectionsPanel />
+
+        {/* 2026-07-17 UX/IA audit M7 — the opt-in nutrients sync inventory
+            (moved from Settings → Sources, where "source priority" only
+            ever meant which-connection-wins ranking, not a read-only sync
+            list). The card carries its own header + renders nothing while
+            the `nutrients` module is off, so it slots in as a sibling card
+            under the same Connections group rather than a new labelled
+            section that could paint an orphaned heading over nothing. */}
+        <NutrientIntakeCard />
      </section>

      {/* Delivery channels ("where it's delivered"). The existing channels
@@ -45,7 +45,6 @@ import { Skeleton } from "@/components/ui/skeleton";
 import { EmptyState } from "@/components/ui/empty-state";
 import { SettingsCard } from "@/components/settings/settings-card";
 import { SettingsCardHeader } from "@/components/settings/_card-header";
-import { NutrientIntakeCard } from "@/components/settings/nutrient-intake-card";
 import { useTranslations } from "@/lib/i18n/context";
 import { queryKeys } from "@/lib/query-keys";
 import {
@@ -585,9 +584,11 @@ export function SourcesSection() {
        </p>
      </SettingsCard>

-      {/* v1.28 — read-only synced-nutrient list for the opt-in `nutrients`
-          module (renders nothing while the module is off). */}
-      <NutrientIntakeCard />
+      {/* v1.28 shipped `<NutrientIntakeCard>` here; the 2026-07-17 UX/IA
+          audit (M7) moved it to Settings → Integrations — it is a read-only
+          Apple-Health sync inventory, not a which-connection-wins ranking,
+          so it never belonged on the "source priority" ladder. See
+          `<IntegrationsSection>` for the card's new home. */}
    </div>
  );
 }
@@ -0,0 +1,22 @@
+"use client";
+
+import { useEffect, useState } from "react";
+
+/**
+ * Debounces a fast-changing value (typically a search-input draft) so a
+ * dependent effect / query key only reacts once the value has settled for
+ * `delayMs`. Mirrors the inline 200 ms pattern the Dokumente vault search
+ * already used (`documents-view.tsx`), centralised here so other search
+ * surfaces (v1.30.2 — the Coach conversation history) don't hand-roll the
+ * same timeout.
+ */
+export function useDebouncedValue<T>(value: T, delayMs = 200): T {
+  const [debounced, setDebounced] = useState(value);
+
+  useEffect(() => {
+    const handle = setTimeout(() => setDebounced(value), delayMs);
+    return () => clearTimeout(handle);
+  }, [value, delayMs]);
+
+  return debounced;
+}
@@ -0,0 +1,59 @@
+"use client";
+
+import { useEffect, useRef, type RefObject } from "react";
+
+/**
+ * IntersectionObserver-driven "load more" sentinel for a scrollable list
+ * backed by `useInfiniteQuery`. Mount the returned ref on an empty element
+ * at the end of the rendered list; once it scrolls into `root` (the list's
+ * own scroll container — pass `null` to watch the viewport instead),
+ * `onLoadMore` fires. Gated by `enabled` so an in-flight fetch or an
+ * exhausted list (no `nextCursor`) never re-triggers.
+ *
+ * v1.30.2 — extracted for the Coach conversation history (rail + the
+ * standalone `/coach/conversations` page), which both need the same
+ * scroll-to-load-more behaviour the Dokumente vault timeline gets from its
+ * `@tanstack/react-virtual` windowing; a virtualizer would be overkill here
+ * (conversation counts don't approach document-vault scale), so this is the
+ * plain-DOM equivalent of that same "pull the next page near the end of the
+ * window" contract.
+ */
+export function useLoadMoreSentinel({
+  enabled,
+  onLoadMore,
+  root = null,
+}: {
+  /** True only when there IS a next page and no fetch is already in flight. */
+  enabled: boolean;
+  onLoadMore: () => void;
+  /** Scroll container to observe within; `null` watches the viewport. */
+  root?: Element | null;
+}): RefObject<HTMLDivElement | null> {
+  const sentinelRef = useRef<HTMLDivElement | null>(null);
+  // Latest-callback-wins ref so the observer callback never captures a stale
+  // `onLoadMore` closure without having to re-create the observer on every
+  // render.
+  const onLoadMoreRef = useRef(onLoadMore);
+  useEffect(() => {
+    onLoadMoreRef.current = onLoadMore;
+  }, [onLoadMore]);
+
+  useEffect(() => {
+    if (!enabled) return;
+    const el = sentinelRef.current;
+    if (!el) return;
+    if (typeof IntersectionObserver === "undefined") return;
+    const observer = new IntersectionObserver(
+      (entries) => {
+        if (entries.some((entry) => entry.isIntersecting)) {
+          onLoadMoreRef.current();
+        }
+      },
+      { root, rootMargin: "200px" },
+    );
+    observer.observe(el);
+    return () => observer.disconnect();
+  }, [enabled, root]);
+
+  return sentinelRef;
+}
@@ -103,6 +103,32 @@ describe("listConversations — filter + DTO", () => {
    expect(where.attachments).toEqual({ some: { documentId: "doc-9" } });
  });

+  // v1.30.2 (QoL H1) — server-side title search for the full-history rail.
+  it("applies a case-insensitive title-contains filter when q is given", async () => {
+    findMany.mockResolvedValue([]);
+    await listConversations({ userId: "user-1", q: "blood pressure" });
+    const where = findMany.mock.calls[0][0].where;
+    expect(where.userId).toBe("user-1");
+    expect(where.title).toEqual({
+      contains: "blood pressure",
+      mode: "insensitive",
+    });
+  });
+
+  it("trims q and omits the title filter for an empty/whitespace-only query", async () => {
+    findMany.mockResolvedValue([]);
+    await listConversations({ userId: "user-1", q: "   " });
+    const where = findMany.mock.calls[0][0].where;
+    expect("title" in where).toBe(false);
+  });
+
+  it("omits the title filter entirely when q is not given (existing behaviour unchanged)", async () => {
+    findMany.mockResolvedValue([]);
+    await listConversations({ userId: "user-1" });
+    const where = findMany.mock.calls[0][0].where;
+    expect("title" in where).toBe(false);
+  });
+
  it("falls back to filename when the first attachment has no title", async () => {
    findMany.mockResolvedValue([
      {
@@ -435,6 +435,16 @@ export interface ListConversationsParams {
   * live join row). `userId` stays narrowed regardless.
   */
  attachedDocumentId?: string;
+  /**
+   * v1.30.2 (QoL H1) — optional server-side title search for the history
+   * rail + the standalone conversations page. Case-insensitive substring
+   * match against `title` only — the only plaintext column on the row;
+   * `CoachMessage.encryptedContent` cannot be searched without decrypting
+   * every message, so message BODIES are out of scope for this pass (see
+   * the route's doc comment). Trimmed empty string is treated as "no
+   * filter", matching the pre-existing cursor/limit handling style.
+   */
+  q?: string;
 }

 /**
@@ -449,12 +459,14 @@ export async function listConversations(
  nextCursor: string | null;
 }> {
  const limit = Math.min(Math.max(params.limit ?? 20, 1), 50);
+  const q = params.q?.trim();
  const rows = await prisma.coachConversation.findMany({
    where: {
      userId: params.userId,
      ...(params.attachedDocumentId
        ? { attachments: { some: { documentId: params.attachedDocumentId } } }
        : {}),
+      ...(q ? { title: { contains: q, mode: "insensitive" } } : {}),
    },
    orderBy: [{ updatedAt: "desc" }, { id: "desc" }],
    take: limit + 1,
@@ -0,0 +1,157 @@
+/**
+ * Unit coverage for `loadIntradayPulse` — the two DATAINT-audit fixes:
+ *
+ *   - M2: the PULSE/step reads are bounded to the exact local-day window
+ *     (`localDayWindow`), not the ±15h/39h padded superset a dense account's
+ *     read cap could be entirely consumed by (the previous evening's
+ *     samples starving the viewed day's afternoon).
+ *   - M5: the resting-baseline PULSE-proxy fallback buckets by the CALLER's
+ *     timezone, not a hardcoded Berlin default — `resolveBaseline` now
+ *     threads `timezone` into `resolveRestingPulseSeries`'s `dayKeyOf`.
+ */
+import { describe, expect, it, vi, beforeEach } from "vitest";
+
+const findManyMock = vi.fn();
+const workoutFindManyMock = vi.fn();
+
+vi.mock("@/lib/db", () => ({
+  prisma: {
+    measurement: { findMany: (...args: unknown[]) => findManyMock(...args) },
+    workout: { findMany: (...args: unknown[]) => workoutFindManyMock(...args) },
+  },
+}));
+
+import { loadIntradayPulse } from "@/lib/analytics/intraday-pulse-io";
+import {
+  resolveRestingPulseSeries,
+  type PulseSample,
+} from "@/lib/analytics/resting-pulse";
+import { toBerlinDayKey } from "@/lib/tz/resolver";
+import { userDayKey } from "@/lib/tz/format";
+import { percentile } from "@/lib/insights/strain-score";
+import { localDayWindow } from "@/lib/measurements/consolidation-tz";
+
+/** Mirrors the private `median()` helper in `intraday-pulse-io.ts`. */
+function median(values: number[]): number | null {
+  if (values.length === 0) return null;
+  return Math.round(percentile(values, 50) * 10) / 10;
+}
+
+describe("loadIntradayPulse", () => {
+  beforeEach(() => {
+    findManyMock.mockReset();
+    workoutFindManyMock.mockReset();
+    workoutFindManyMock.mockResolvedValue([]);
+  });
+
+  it("M2 — queries PULSE/step rows against the exact local-day window, not the padded lead/trail superset", async () => {
+    const tz = "America/New_York";
+    const dateKey = "2026-06-15";
+    const { dayStart, dayEnd } = localDayWindow(dateKey, tz);
+
+    findManyMock.mockResolvedValue([]);
+
+    await loadIntradayPulse("user-1", tz, dateKey);
+
+    const pulseCall = findManyMock.mock.calls.find(
+      (call) => call[0]?.where?.type === "PULSE",
+    );
+    const stepCall = findManyMock.mock.calls.find(
+      (call) => call[0]?.where?.type === "ACTIVITY_STEPS",
+    );
+    expect(pulseCall).toBeDefined();
+    expect(stepCall).toBeDefined();
+    expect(pulseCall![0].where.measuredAt).toEqual({
+      gte: dayStart,
+      lt: dayEnd,
+    });
+    expect(stepCall![0].where.measuredAt).toEqual({
+      gte: dayStart,
+      lt: dayEnd,
+    });
+  });
+
+  it("M5 — the PULSE-proxy baseline buckets by the caller's timezone, not Berlin", async () => {
+    const tz = "Pacific/Kiritimati"; // UTC+14, no DST — deterministic offset
+    const dateKey = "2026-06-15";
+
+    // Two UTC instants that land on DIFFERENT Kiritimati calendar days but
+    // the SAME Berlin calendar day (Kiritimati is ~13h ahead of Berlin in
+    // June) — a clean tz-bucketing fork.
+    const dayA = [
+      new Date("2026-06-10T09:00:00.000Z"), // Kiritimati: 2026-06-10T23:00 (day A)
+      new Date("2026-06-10T10:00:00.000Z"),
+      new Date("2026-06-10T11:00:00.000Z"),
+    ];
+    const dayB = [
+      new Date("2026-06-10T12:00:00.000Z"), // Kiritimati: 2026-06-11T02:00 (day B)
+      new Date("2026-06-10T13:00:00.000Z"),
+      new Date("2026-06-10T14:00:00.000Z"),
+    ];
+    expect(userDayKey(dayA[0], tz)).not.toBe(userDayKey(dayB[0], tz));
+    // Sanity: under the OLD hardcoded-Berlin bug all six samples fall on the
+    // SAME Berlin day — proving the two tz groupings genuinely diverge.
+    const berlinDays = new Set(
+      [...dayA, ...dayB].map((d) => toBerlinDayKey(d)),
+    );
+    expect(berlinDays.size).toBe(1);
+
+    const pulseHistory: Array<{ value: number; measuredAt: Date }> = [
+      ...dayA.map((measuredAt, i) => ({ value: 50 + i, measuredAt })),
+      ...dayB.map((measuredAt, i) => ({ value: 80 + i, measuredAt })),
+    ];
+
+    findManyMock.mockImplementation(
+      async (args: { where: { type: string } }) => {
+        if (args.where.type === "PULSE") return [];
+        if (args.where.type === "RESTING_HEART_RATE") return [];
+        return [];
+      },
+    );
+    // `resolveBaseline` fires a SECOND PULSE read (take: 365, no window) when
+    // resting rows are thin — swap the mock to serve it on the second PULSE
+    // call specifically by checking call order isn't reliable, so key off
+    // `select`/`take` shape instead.
+    findManyMock.mockImplementation(
+      async (args: {
+        where: { type: string; measuredAt?: unknown };
+        take?: number;
+      }) => {
+        if (args.where.type === "RESTING_HEART_RATE") return [];
+        if (args.where.type === "PULSE" && args.take === 365) {
+          return pulseHistory;
+        }
+        // The view-day-windowed PULSE / ACTIVITY_STEPS reads — irrelevant here.
+        return [];
+      },
+    );
+
+    const result = await loadIntradayPulse("user-1", tz, dateKey);
+
+    expect(result.baselineSource).toBe("proxy");
+
+    const expected = resolveRestingPulseSeries({
+      restingSamples: [],
+      pulseSamples: pulseHistory as PulseSample[],
+      dayKeyOf: (d) => userDayKey(d, tz),
+    });
+    const expectedBaseline = median(
+      expected.series.slice(-30).map((p) => p.value),
+    );
+    expect(result.baseline).toBe(expectedBaseline);
+
+    // Regression guard: the OLD hardcoded-Berlin bucketing would have folded
+    // every sample into ONE day (see the `berlinDays.size === 1` sanity
+    // check above) and produced a materially different proxy value — assert
+    // the fixed result does NOT match that wrong grouping's outcome.
+    const buggyBerlin = resolveRestingPulseSeries({
+      restingSamples: [],
+      pulseSamples: pulseHistory as PulseSample[],
+      dayKeyOf: (d) => toBerlinDayKey(d),
+    });
+    const buggyBaseline = median(
+      buggyBerlin.series.slice(-30).map((p) => p.value),
+    );
+    expect(result.baseline).not.toBe(buggyBaseline);
+  });
+});
@@ -18,6 +18,8 @@
 import { prisma } from "@/lib/db";
 import { percentile } from "@/lib/insights/strain-score";
 import { resolveRestingPulseSeries } from "@/lib/analytics/resting-pulse";
+import { localDayWindow } from "@/lib/measurements/consolidation-tz";
+import { userDayKey } from "@/lib/tz/format";
 import {
  BUCKET_MINUTES,
  HOURLY_BUCKET_MINUTES,
@@ -73,8 +75,18 @@ function median(values: number[]): number | null {
 * baseline is the median of the resolved daily series — robust to the odd
 * outlier — and is "mature" only once at least `MIN_BASELINE_DAYS` distinct
 * daily points exist.
+ *
+ * `timezone` buckets the PULSE-proxy fallback in the USER's local day
+ * (DATAINT M5) — the caller previously omitted it, so
+ * `deriveRestingProxyFromPulse` defaulted to Berlin-day buckets for every
+ * user. A non-Berlin user's late-evening (often workout-heavy) samples then
+ * folded into the wrong day's bucket, shifting the daily floor the tension
+ * detector's `baseline + 12 bpm` threshold judges the day against.
 */
-async function resolveBaseline(userId: string): Promise<{
+async function resolveBaseline(
+  userId: string,
+  timezone: string,
+): Promise<{
  baseline: number | null;
  mature: boolean;
  source: "resting" | "proxy" | "none";
@@ -106,6 +118,7 @@ async function resolveBaseline(userId: string): Promise<{
      measuredAt: r.measuredAt,
      value: r.value,
    })),
+    dayKeyOf: (d) => userDayKey(d, timezone),
  });

  const recent = resolved.series.slice(-30).map((p) => p.value);
@@ -148,6 +161,18 @@ export async function loadIntradayPulse(
  const anchorMs = new Date(`${dateKey}T00:00:00.000Z`).getTime();
  const start = new Date(anchorMs - WINDOW_LEAD_MS);
  const end = new Date(anchorMs + WINDOW_TRAIL_MS);
+  // DATAINT M2 — PULSE/step reads are bounded to the EXACT local-day window
+  // (DST-aware; 23/24/25h), not the ±15h/39h padded superset. The prior
+  // padded read meant `take: MAX_DAY_*_ROWS` (an ascending cap) could be
+  // entirely consumed by the PREVIOUS local day's samples (e.g. a
+  // workout-heavy evening) before reaching the viewed day at all on a dense
+  // account — the cap now only ever spends against rows that are actually
+  // inside `dateKey`. `computeTenMinuteMeanSeries` / the step-bucket loop
+  // already re-filter by `dayKey` defensively, so this is a pure narrowing.
+  // The workout-overlap read keeps the wider padded window: it has no `take`
+  // cap, so there is no truncation risk, and a workout can legitimately
+  // start the local-day before or end the local-day after.
+  const { dayStart, dayEnd } = localDayWindow(dateKey, timezone);

  const [pulseRows, stepRows, workoutRows, baseline] = await Promise.all([
    prisma.measurement.findMany({
@@ -155,7 +180,7 @@ export async function loadIntradayPulse(
        userId,
        type: "PULSE",
        deletedAt: null,
-        measuredAt: { gte: start, lte: end },
+        measuredAt: { gte: dayStart, lt: dayEnd },
      },
      orderBy: { measuredAt: "asc" },
      take: MAX_DAY_PULSE_ROWS,
@@ -166,7 +191,7 @@ export async function loadIntradayPulse(
        userId,
        type: "ACTIVITY_STEPS",
        deletedAt: null,
-        measuredAt: { gte: start, lte: end },
+        measuredAt: { gte: dayStart, lt: dayEnd },
      },
      orderBy: { measuredAt: "asc" },
      take: MAX_DAY_STEP_ROWS,
@@ -180,7 +205,7 @@ export async function loadIntradayPulse(
      },
      select: { startedAt: true, endedAt: true },
    }),
-    resolveBaseline(userId),
+    resolveBaseline(userId, timezone),
  ]);

  // Split live PULSE rows into raw per-sample rows and already-folded
@@ -335,7 +335,14 @@ describe("buildDailyDigest — coach check-in (S3)", () => {
    expect(item?.body).toContain("every morning → weigh in");
    // Keep / let-go carry the plan id; adjust is a plain navigation href.
    expect(item?.actions[0].intent).toBe(`${COACH_CHECKIN_KEEP_INTENT}:p1`);
-    expect(item?.actions[1].href).toBe("/coach");
+    // 2026-07-17 UX-flows audit F1-2 — adjust used to be a bare `/coach` link
+    // dropping all plan context; it now carries `?ask=` seeding the coach
+    // composer with the plan's own words, so the target is the coach route
+    // with the plan text echoed in the query, not a blank chat.
+    expect(item?.actions[1].href).toContain("/coach?ask=");
+    expect(item?.actions[1].href).toContain(
+      encodeURIComponent("every morning → weigh in"),
+    );
    expect(item?.actions[2].intent).toBe(`${COACH_CHECKIN_LETGO_INTENT}:p1`);
  });

@@ -451,6 +451,18 @@ function buildCoachCheckinItem(
    ? t("daily.item.coachCheckin.bodyPlan", { plan: plan.planText })
    : t("daily.item.coachCheckin.body");

+  // 2026-07-17 UX-flows audit F1-2 — "Adjust" used to be a bare `/coach`
+  // link carrying no plan context: keep/let-go both thread the plan id
+  // into their intent, but adjust dropped the user into an empty new chat
+  // with no memory of which plan prompted the tap. The coach route's
+  // full-page surface now reads `?ask=` and seeds it as the composer
+  // prefill for a fresh conversation (`CoachConversation`'s `prefill` prop,
+  // same mechanism the drawer's suggested-prompt chips use) — the user
+  // lands with the plan already named instead of a blank composer.
+  const adjustPrompt = plan.planText
+    ? t("daily.item.coachCheckin.adjustPromptPlan", { plan: plan.planText })
+    : t("daily.item.coachCheckin.adjustPrompt");
+
  return {
    kind: "coach_checkin",
    title: t("daily.item.coachCheckin.title"),
@@ -464,7 +476,7 @@ function buildCoachCheckinItem(
      {
        labelKey: "daily.action.checkinAdjust",
        intent: COACH_CHECKIN_ADJUST_INTENT,
-        href: "/coach",
+        href: `/coach?ask=${encodeURIComponent(adjustPrompt)}`,
      },
      {
        labelKey: "daily.action.checkinLetGo",
@@ -73,6 +73,22 @@ describe("getMissFreeDayKeys", () => {
    ]);
    expect(keys).toEqual(["2026-05-01"]);
  });
+
+  it("buckets by the CALLER's timezone when supplied (DATAINT M4)", () => {
+    // 2026-05-01T23:30:00Z: Berlin (CEST, UTC+2) reads 2026-05-02T01:30 —
+    // the NEXT day. A resolved slot at that instant must key to 05-01 for
+    // an America/New_York (UTC-4 in May) caller, not 05-02.
+    const lateEvent: CareIntakeEventRecord = {
+      scheduledFor: new Date("2026-05-01T23:30:00Z"),
+      takenAt: new Date("2026-05-01T23:35:00Z"),
+      skipped: false,
+      autoMissed: false,
+    };
+    const berlinDefault = getMissFreeDayKeys([lateEvent]);
+    const newYork = getMissFreeDayKeys([lateEvent], "America/New_York");
+    expect(berlinDefault).toEqual(["2026-05-02"]);
+    expect(newYork).toEqual(["2026-05-01"]);
+  });
 });

 describe("getWeeklyConsistency", () => {
@@ -268,6 +268,61 @@ describe("getHiddenMetrics", () => {
  });
 });

+describe("getEngagementMetrics — tz threading (DATAINT M4)", () => {
+  it("buckets entry-day-streak by the CALLER's timezone, not a hardcoded Berlin default", () => {
+    // Two instants 2.5h apart in UTC. Under Berlin (CEST, UTC+2 in June)
+    // they straddle a MIDNIGHT boundary and land on two DIFFERENT
+    // consecutive local days (streak 2). Under Kiritimati (UTC+14) both
+    // fall on the SAME local day (streak 1). If `tz` were silently ignored
+    // (still hardcoded Berlin), both calls would report streak 2.
+    const measurements = [
+      { type: "WEIGHT", measuredAt: new Date("2026-06-10T21:00:00.000Z") },
+      { type: "PULSE", measuredAt: new Date("2026-06-10T23:30:00.000Z") },
+    ];
+    const berlinDefault = getEngagementMetrics({
+      measurements,
+      moodEntries: [],
+      intakeEvents: [],
+    });
+    const kiritimati = getEngagementMetrics({
+      measurements,
+      moodEntries: [],
+      intakeEvents: [],
+      tz: "Pacific/Kiritimati",
+    });
+    expect(berlinDefault.entryDayStreak).toBe(2);
+    expect(kiritimati.entryDayStreak).toBe(1);
+  });
+});
+
+describe("getHiddenMetrics — tz threading (DATAINT M4)", () => {
+  it("classifies night-owl/early-bird hours against the caller's own timezone", () => {
+    // 2026-04-15T00:30:00Z: Berlin (CEST, UTC+2) reads 02:30 → night-owl.
+    // The same instant in America/Los_Angeles (PDT, UTC-7) reads 17:30 the
+    // PREVIOUS day → neither night-owl nor early-bird. If `tz` were still
+    // hardcoded Berlin, both calls would report the same night-owl count.
+    const measurements = [
+      { type: "WEIGHT", measuredAt: new Date("2026-04-15T00:30:00Z") },
+    ];
+    const berlin = getHiddenMetrics({
+      measurements,
+      moodEntries: [],
+      intakeEvents: [],
+      auditEvents: [],
+    });
+    const losAngeles = getHiddenMetrics({
+      measurements,
+      moodEntries: [],
+      intakeEvents: [],
+      auditEvents: [],
+      tz: "America/Los_Angeles",
+    });
+    expect(berlin.nightOwlCount).toBe(1);
+    expect(losAngeles.nightOwlCount).toBe(0);
+    expect(losAngeles.earlyBirdCount).toBe(0);
+  });
+});
+
 describe("getEarnabilityFlags", () => {
  it("flags categories the user has data for", () => {
    expect(
@@ -22,7 +22,6 @@ import {
  calculateLongestStreak,
  evaluateAchievementsWithCompletionDates,
  moduleForMetric,
-  toBerlinDayKey,
  type AchievementMetrics,
  type AchievementProgress,
 } from "@/lib/gamification/achievements";
@@ -41,6 +40,7 @@ import {
 } from "@/lib/analytics/classifications";
 import { classifyIntakeTiming } from "@/lib/analytics/compliance";
 import { wallClockInTz } from "@/lib/tz/wall-clock";
+import { DEFAULT_TIMEZONE, userDayKey } from "@/lib/tz/format";

 export type AuthedUser = Awaited<ReturnType<typeof requireAuth>>["user"];

@@ -74,22 +74,23 @@ type MedicationScheduleRecord = {

 /**
 * v1.28.25 — one SQL-aggregated vitals bucket: every non-deleted, non-import
- * row of `type` inside the Berlin-local hour `hour` of day `dayKey`, folded
- * to a row count + value sum by the database. The vitals window is
- * rollout-date → now and grows forever, and PULSE can be per-sample heart
- * rate (six figures on a long-tenured wearable account) — reading raw rows
- * here was the scale hazard. Every consumer's semantics survive the fold:
- * green-day classification needs the per-day per-type MEAN (= Σsum / Σcount
- * across the day's hour buckets — identical to the raw mean up to
- * floating-point summation order), the count badges need row counts, the
- * engagement / consistency series need day presence, and the hidden
+ * row of `type` inside the badge-owner's LOCAL hour `hour` of day `dayKey`
+ * (DATAINT M4 — was hardcoded Berlin; the query now binds the caller's own
+ * `User.timezone`), folded to a row count + value sum by the database. The
+ * vitals window is rollout-date → now and grows forever, and PULSE can be
+ * per-sample heart rate (six figures on a long-tenured wearable account) —
+ * reading raw rows here was the scale hazard. Every consumer's semantics
+ * survive the fold: green-day classification needs the per-day per-type MEAN
+ * (= Σsum / Σcount across the day's hour buckets — identical to the raw mean
+ * up to floating-point summation order), the count badges need row counts,
+ * the engagement / consistency series need day presence, and the hidden
 * night-owl / early-bird / leap-day tallies need per-hour row counts.
 */
 type VitalsHourBucket = {
  type: "WEIGHT" | "BLOOD_PRESSURE_SYS" | "BLOOD_PRESSURE_DIA" | "PULSE";
-  /** Berlin-local YYYY-MM-DD. */
+  /** User-local YYYY-MM-DD. */
  dayKey: string;
-  /** Berlin-local hour 0–23. */
+  /** User-local hour 0–23. */
  hour: number;
  /** Number of raw rows folded into this bucket. */
  rowCount: number;
@@ -102,21 +103,22 @@ function maxDate(a: Date, b: Date): Date {
 }

 /**
- * v1.18.1 P4 — every Berlin-local day key covered by an illness episode
- * (onset day through the resolved day, or through `now` while ongoing),
- * inclusive on both ends. These are the days a streak is allowed to lapse
- * across without breaking (the Rest Mode freeze). Uses the same
- * `toBerlinDayKey` day anchoring every other streak series uses, so the
+ * v1.18.1 P4 — every local day key covered by an illness episode (onset day
+ * through the resolved day, or through `now` while ongoing), inclusive on
+ * both ends. These are the days a streak is allowed to lapse across without
+ * breaking (the Rest Mode freeze). Uses the same `tz` day anchoring every
+ * other streak series uses (DATAINT M4 — was hardcoded Berlin), so the
 * frozen days line up with the qualifying days exactly.
 */
 function collectIllnessDayKeys(
  episodes: Array<{ onsetAt: Date; resolvedAt: Date | null }>,
  now: Date,
+  tz: string,
 ): Set<string> {
  const frozen = new Set<string>();
  for (const ep of episodes) {
-    const startSerial = dayKeyToSerial(toBerlinDayKey(ep.onsetAt));
-    const endSerial = dayKeyToSerial(toBerlinDayKey(ep.resolvedAt ?? now));
+    const startSerial = dayKeyToSerial(userDayKey(ep.onsetAt, tz));
+    const endSerial = dayKeyToSerial(userDayKey(ep.resolvedAt ?? now, tz));
    for (let serial = startSerial; serial <= endSerial; serial++) {
      frozen.add(serialToDayKey(serial));
    }
@@ -188,14 +190,17 @@ function toDaySeries(dayKeys: string[]): {
  };
 }

-function getEventDaySeries(dates: Date[]): {
+function getEventDaySeries(
+  dates: Date[],
+  tz: string,
+): {
  dayKeys: string[];
  firstDateByDay: Map<string, Date>;
 } {
  const firstDateByDay = new Map<string, Date>();

  for (const date of dates) {
-    const dayKey = toBerlinDayKey(date);
+    const dayKey = userDayKey(date, tz);
    const existing = firstDateByDay.get(dayKey);

    if (!existing || date < existing) {
@@ -288,7 +293,10 @@ function getOnTimePerfectDaySeries(
  const eventsByDay = new Map<string, IntakeEventRecord[]>();

  for (const event of intakeEvents) {
-    const dayKey = toBerlinDayKey(event.scheduledFor);
+    // DATAINT M4 — was `toBerlinDayKey`; this function already threads `tz`
+    // for the schedule-window classification below, but the day-BUCKETING
+    // key itself was still hardcoded Berlin.
+    const dayKey = userDayKey(event.scheduledFor, tz);
    const list = eventsByDay.get(dayKey) ?? [];
    list.push(event);
    eventsByDay.set(dayKey, list);
@@ -392,6 +400,7 @@ function getExpectedIntakesForDay(
 function getIntakeIssueMetrics(
  intakeEvents: IntakeEventRecord[],
  schedulesByMedicationId: Map<string, MedicationScheduleRecord[]>,
+  tz: string,
 ) {
  const skippedIntakeDates = intakeEvents
    .filter((event) => event.skipped)
@@ -408,7 +417,7 @@ function getIntakeIssueMetrics(
      continue;
    }

-    const dayKey = toBerlinDayKey(event.scheduledFor);
+    const dayKey = userDayKey(event.scheduledFor, tz);
    const byDay = eventsByMedicationDay.get(event.medicationId) ?? new Map();
    const list = byDay.get(dayKey) ?? [];
    list.push(event);
@@ -470,11 +479,12 @@ function getCompliance80DaySeries(
  intakeEvents: IntakeEventRecord[],
  startDate: Date,
  endDate: Date,
+  tz: string,
 ) {
  const perDayCounts = new Map<number, { logged: number; taken: number }>();

  for (const event of intakeEvents) {
-    const serial = dayKeyToSerial(toBerlinDayKey(event.scheduledFor));
+    const serial = dayKeyToSerial(userDayKey(event.scheduledFor, tz));
    const current = perDayCounts.get(serial) ?? { logged: 0, taken: 0 };

    const logged = event.skipped || event.takenAt !== null;
@@ -486,8 +496,8 @@ function getCompliance80DaySeries(
    perDayCounts.set(serial, current);
  }

-  const startSerial = dayKeyToSerial(toBerlinDayKey(startDate));
-  const endSerial = dayKeyToSerial(toBerlinDayKey(endDate));
+  const startSerial = dayKeyToSerial(userDayKey(startDate, tz));
+  const endSerial = dayKeyToSerial(userDayKey(endDate, tz));

  let rollingLogged = 0;
  let rollingTaken = 0;
@@ -538,6 +548,14 @@ export async function buildAchievementsResult(
  const now = new Date();
  const userId = user.id;
  const startDate = maxDate(user.createdAt, GAMIFICATION_ROLLOUT_AT);
+  // v1.30 (DATAINT M4) — every day-key derivation below anchors on the
+  // BADGE OWNER's own timezone, not a hardcoded Berlin default. Before this,
+  // a non-CET user's evening entries could fold into the wrong calendar day
+  // (or merge two real local days into one), silently breaking every
+  // day-streak badge (login, entry, weekend, miss-free, sleep-log, on-time,
+  // compliance-80) plus the hidden night-owl / early-bird / leap-day
+  // triggers for every non-CET account.
+  const tz = user.timezone || DEFAULT_TIMEZONE;

  const [
    vitalsBuckets,
@@ -550,7 +568,7 @@ export async function buildAchievementsResult(
    healthProfile,
    illnessEpisodes,
  ] = await Promise.all([
-    // v1.28.25 — vitals folded to (type, Berlin day, Berlin hour) buckets in
+    // v1.28.25 — vitals folded to (type, local day, local hour) buckets in
    // SQL instead of one row per sample. The window (rollout → now) grows
    // forever and PULSE can be per-sample heart rate, so the raw read ran to
    // six figures on long-tenured wearable accounts; the bucket set is
@@ -561,18 +579,21 @@ export async function buildAchievementsResult(
    // (never user input), source ≠ IMPORT, and soft-deleted rows stay
    // excluded (v1.4.41 W-DELETED-2) so deleting a row immediately rolls
    // back any streak / count badge it earned. `AT TIME ZONE 'UTC' AT TIME
-    // ZONE 'Europe/Berlin'` re-tags the naive UTC `measured_at` and converts
-    // it to the same Berlin wall clock `toBerlinDayKey` / `berlinHour`
-    // derive via Intl. Ordering pins day → hour → type for determinism.
+    // ZONE ${tz}` re-tags the naive UTC `measured_at` and converts it to the
+    // BADGE OWNER's own wall clock (DATAINT M4 — was hardcoded
+    // 'Europe/Berlin'; `tz` is bound as a query parameter, never spliced, and
+    // is only ever a value `PUT /api/auth/me/timezone` already validated
+    // against `Intl`'s IANA zone list before it reached the column).
+    // Ordering pins day → hour → type for determinism.
    prisma.$queryRaw<VitalsHourBucket[]>`
      SELECT
        m."type"::text AS "type",
        to_char(
-          (m."measured_at" AT TIME ZONE 'UTC') AT TIME ZONE 'Europe/Berlin',
+          (m."measured_at" AT TIME ZONE 'UTC') AT TIME ZONE ${tz},
          'YYYY-MM-DD'
        ) AS "dayKey",
        EXTRACT(
-          HOUR FROM (m."measured_at" AT TIME ZONE 'UTC') AT TIME ZONE 'Europe/Berlin'
+          HOUR FROM (m."measured_at" AT TIME ZONE 'UTC') AT TIME ZONE ${tz}
        )::int AS "hour",
        COUNT(*)::int AS "rowCount",
        SUM(m."value")::double precision AS "valueSum"
@@ -708,12 +729,12 @@ export async function buildAchievementsResult(
      : Promise.resolve([]),
  ]);

-  // v1.18.1 P4 — Rest Mode streak-freeze. Build the set of Berlin-local day
-  // keys covered by an illness episode (onset → resolved, or → now while
+  // v1.18.1 P4 — Rest Mode streak-freeze. Build the set of local-day keys
+  // covered by an illness episode (onset → resolved, or → now while
  // ongoing). A day-streak that lapses only across these days is bridged
  // rather than broken (see `bridgeFrozenStreakGaps`): being unwell pauses a
  // streak, it does not fail it.
-  const frozenIllnessDayKeys = collectIllnessDayKeys(illnessEpisodes, now);
+  const frozenIllnessDayKeys = collectIllnessDayKeys(illnessEpisodes, now, tz);
  const freeze = (dayKeys: string[]): string[] =>
    bridgeFrozenStreakGaps(dayKeys, frozenIllnessDayKeys);

@@ -723,6 +744,7 @@ export async function buildAchievementsResult(
  const intakeIssueMetrics = getIntakeIssueMetrics(
    intakeEvents,
    schedulesByMedicationId,
+    tz,
  );

  const takenIntakeDates = intakeEvents
@@ -735,12 +757,12 @@ export async function buildAchievementsResult(
  const passwordLoginDates = auditEvents
    .filter((event) => event.action === "auth.login.password")
    .map((event) => event.createdAt);
-  const loginDaySeries = getEventDaySeries([
-    ...passkeyLoginDates,
-    ...passwordLoginDates,
-  ]);
+  const loginDaySeries = getEventDaySeries(
+    [...passkeyLoginDates, ...passwordLoginDates],
+    tz,
+  );

-  // v1.28.25 — the buckets already carry the Berlin day-key + hour straight
+  // v1.28.25 — the buckets already carry the local day-key + hour straight
  // from SQL (superseding the v1.18.11 Intl-once optimisation: no Intl pass
  // over the vitals remains at all). The parallel arrays thread into every
  // consumer exactly as the per-row keys did; day-presence consumers dedup
@@ -762,21 +784,23 @@ export async function buildAchievementsResult(
  const onTimeSeries = getOnTimePerfectDaySeries(
    intakeEvents,
    schedulesByMedicationId,
-    user.timezone,
+    tz,
  );
  const complianceSeries = getCompliance80DaySeries(
    intakeEvents,
    startDate,
    now,
+    tz,
  );

  // v1.16.1 — care-routine series. Miss-free days and sleep-logging
  // days reuse the standard day-series plumbing; the weekly measurement
  // consistency folds distinct active vitals days into Monday-anchored
  // weeks (≥5 active days each, 4 consecutive weeks for the badge).
-  const missFreeSeries = toDaySeries(getMissFreeDayKeys(intakeEvents));
+  const missFreeSeries = toDaySeries(getMissFreeDayKeys(intakeEvents, tz));
  const sleepSeries = getEventDaySeries(
    sleepMeasurements.map((m) => m.measuredAt),
+    tz,
  );
  const weeklyConsistency = getWeeklyConsistency(
    // Bucket day-keys — `getWeeklyConsistency` dedups to distinct days, so
@@ -799,11 +823,12 @@ export async function buildAchievementsResult(
    moodEntries,
    intakeEvents,
    auditEvents,
-    // The SQL-derived Berlin day-keys + hours ride in parallel to the
+    // The SQL-derived local day-keys + hours ride in parallel to the
    // weighted bucket records — the expansion passes never touch a bucket's
    // representative `measuredAt`.
    measurementDayKeys: vitalsDayKeys,
    measurementHours: vitalsHours,
+    tz,
  });
  const earnability = getEarnabilityFlags({
    hasMedication: medications.length > 0,
@@ -24,7 +24,7 @@
 *     boring, clinically useful habit of measuring most days of the
 *     week, week after week, without demanding a perfect daily streak.
 */
-import { toBerlinDayKey } from "./achievements";
+import { DEFAULT_TIMEZONE, userDayKey } from "@/lib/tz/format";

 export interface CareIntakeEventRecord {
  scheduledFor: Date;
@@ -37,11 +37,19 @@ export interface CareIntakeEventRecord {
 * Qualifying day keys for the miss-free streak, sorted ascending.
 * Feed the result to `calculateLongestStreak` /
 * `findStreakCompletionDate` like every other day series.
+ *
+ * `tz` defaults to `DEFAULT_TIMEZONE` ("Europe/Berlin") so every existing
+ * caller/fixture that never set one keeps its prior behaviour byte-for-
+ * byte (DATAINT M4) — the achievements builder now threads the badge
+ * owner's real `User.timezone` in.
 */
-export function getMissFreeDayKeys(events: CareIntakeEventRecord[]): string[] {
+export function getMissFreeDayKeys(
+  events: CareIntakeEventRecord[],
+  tz: string = DEFAULT_TIMEZONE,
+): string[] {
  const byDay = new Map<string, { resolved: number; missed: number }>();
  for (const event of events) {
-    const dayKey = toBerlinDayKey(event.scheduledFor);
+    const dayKey = userDayKey(event.scheduledFor, tz);
    const bucket = byDay.get(dayKey) ?? { resolved: 0, missed: 0 };
    if (event.autoMissed) {
      bucket.missed += 1;
@@ -7,35 +7,27 @@
 * `evaluateAchievementsWithCompletionDates` evaluates the new
 * achievements with no further plumbing.
 */
-import {
-  calculateLongestStreak,
-  toBerlinDayKey,
-  type EarnabilityFlags,
-} from "./achievements";
+import { calculateLongestStreak, type EarnabilityFlags } from "./achievements";
+import { DEFAULT_TIMEZONE, userDayKey } from "@/lib/tz/format";
+import { wallClockInTz } from "@/lib/tz/wall-clock";

-const BERLIN_TIMEZONE = "Europe/Berlin";
-
-const BERLIN_HOUR_FORMATTER = new Intl.DateTimeFormat("en-US", {
-  timeZone: BERLIN_TIMEZONE,
-  hour: "2-digit",
-  hour12: false,
-});
-
-function berlinHour(date: Date): number {
-  const value = BERLIN_HOUR_FORMATTER.format(date);
-  // formatToParts is more reliable across runtimes — split + parseInt
-  // works because the formatter uses 24-hour with no extra tokens.
-  const parsed = Number.parseInt(value, 10);
-  return Number.isNaN(parsed) ? 0 : parsed;
+/**
+ * v1.30 (DATAINT M4) — day-key / hour derivation now takes the CALLER's
+ * timezone (`User.timezone`), defaulting to `DEFAULT_TIMEZONE` ("Europe/
+ * Berlin") when omitted so every existing fixture/caller that never set a
+ * `tz` keeps its prior (Berlin-anchored) behaviour byte-for-byte. Before
+ * this, every derivation hardcoded Berlin regardless of the achievement
+ * owner's real timezone — a non-CET user's evening entries could fold
+ * into the wrong calendar day (or merge two real days into one), silently
+ * breaking day-streak badges (`entryDayStreak`, `weekendStreakCount`) and
+ * the hidden night-owl / early-bird triggers for every non-CET account.
+ */
+function hourInTz(date: Date, tz: string): number {
+  return wallClockInTz(date, tz).hour;
 }

-const BERLIN_WEEKDAY_FORMATTER = new Intl.DateTimeFormat("en-US", {
-  timeZone: BERLIN_TIMEZONE,
-  weekday: "short",
-});
-
-function berlinWeekday(date: Date): string {
-  return BERLIN_WEEKDAY_FORMATTER.format(date);
+function isSaturdayInTz(date: Date, tz: string): boolean {
+  return wallClockInTz(date, tz).weekday === 6;
 }

 export interface MoodEntryRecord {
@@ -165,30 +157,33 @@ export function getEngagementMetrics(input: {
  moodEntries: MoodEntryRecord[];
  intakeEvents: IntakeEventRecord[];
  /**
-   * v1.18.11 (W5 perf) — Berlin day-keys for `measurements`, precomputed
-   * once by the caller and passed in parallel to the array. When present the
-   * measurement rows reuse these instead of re-deriving a `toBerlinDayKey`
-   * per row; mood / intake timestamps (far fewer) still derive inline. Falls
-   * back to per-row derivation when absent — byte-identical result either way.
+   * v1.18.11 (W5 perf) — day-keys for `measurements`, precomputed once by
+   * the caller and passed in parallel to the array. When present the
+   * measurement rows reuse these instead of re-deriving a day key per row;
+   * mood / intake timestamps (far fewer) still derive inline. Falls back to
+   * per-row derivation when absent — byte-identical result either way.
   */
  measurementDayKeys?: string[];
+  /** v1.30 (DATAINT M4) — day-key timezone; defaults to Berlin (see file docblock). */
+  tz?: string;
 }): {
  consistentMonthCount: number;
  entryDayStreak: number;
  weekendStreakCount: number;
 } {
+  const tz = input.tz ?? DEFAULT_TIMEZONE;
  const dayKeyAccum: string[] = [];
  for (let i = 0; i < input.measurements.length; i++) {
    dayKeyAccum.push(
      input.measurementDayKeys?.[i] ??
-        toBerlinDayKey(input.measurements[i].measuredAt),
+        userDayKey(input.measurements[i].measuredAt, tz),
    );
  }
  for (const e of input.moodEntries)
-    dayKeyAccum.push(toBerlinDayKey(e.moodLoggedAt));
+    dayKeyAccum.push(userDayKey(e.moodLoggedAt, tz));
  for (const i of input.intakeEvents) {
-    if (i.takenAt) dayKeyAccum.push(toBerlinDayKey(i.takenAt));
-    else if (i.skipped) dayKeyAccum.push(toBerlinDayKey(i.scheduledFor));
+    if (i.takenAt) dayKeyAccum.push(userDayKey(i.takenAt, tz));
+    else if (i.skipped) dayKeyAccum.push(userDayKey(i.scheduledFor, tz));
  }

  if (dayKeyAccum.length === 0) {
@@ -238,20 +233,20 @@ export function getEngagementMetrics(input: {
  let weekendStreakCount = 0;
  let weekendRun = 0;
  // Walk one Saturday at a time. Use a simple `Date` cursor; the
-  // Berlin TZ formatter resolves the weekday so DST shifts are safe.
+  // caller's tz formatter resolves the weekday so DST shifts are safe.
  const cursor = parseDayKey(firstDay);
  const end = parseDayKey(lastDay);
  // Snap cursor to the next Saturday at 12:00 UTC (avoids DST edges).
  cursor.setUTCHours(12, 0, 0, 0);
-  while (berlinWeekday(cursor) !== "Sat") {
+  while (!isSaturdayInTz(cursor, tz)) {
    cursor.setUTCDate(cursor.getUTCDate() + 1);
    if (cursor > end) break;
  }
  while (cursor <= end) {
-    const sat = toBerlinDayKey(cursor);
+    const sat = userDayKey(cursor, tz);
    const sunDate = new Date(cursor);
    sunDate.setUTCDate(sunDate.getUTCDate() + 1);
-    const sun = toBerlinDayKey(sunDate);
+    const sun = userDayKey(sunDate, tz);
    if (daySet.has(sat) && daySet.has(sun)) {
      weekendRun += 1;
      weekendStreakCount = Math.max(weekendStreakCount, weekendRun);
@@ -290,6 +285,8 @@ export function getHiddenMetrics(input: {
   */
  measurementDayKeys?: string[];
  measurementHours?: number[];
+  /** v1.30 (DATAINT M4) — day-key timezone; defaults to Berlin (see file docblock). */
+  tz?: string;
 }): {
  nightOwlCount: number;
  earlyBirdCount: number;
@@ -297,6 +294,7 @@ export function getHiddenMetrics(input: {
  doctorPdfCount: number;
  localeFlipCount: number;
 } {
+  const tz = input.tz ?? DEFAULT_TIMEZONE;
  let nightOwl = 0;
  let earlyBird = 0;
  let leapDay = 0;
@@ -314,16 +312,16 @@ export function getHiddenMetrics(input: {
  for (let i = 0; i < input.measurements.length; i++) {
    const m = input.measurements[i];
    tally(
-      input.measurementHours?.[i] ?? berlinHour(m.measuredAt),
-      input.measurementDayKeys?.[i] ?? toBerlinDayKey(m.measuredAt),
+      input.measurementHours?.[i] ?? hourInTz(m.measuredAt, tz),
+      input.measurementDayKeys?.[i] ?? userDayKey(m.measuredAt, tz),
      m.count ?? 1,
    );
  }
  for (const e of input.moodEntries) {
-    tally(berlinHour(e.moodLoggedAt), toBerlinDayKey(e.moodLoggedAt));
+    tally(hourInTz(e.moodLoggedAt, tz), userDayKey(e.moodLoggedAt, tz));
  }
  for (const i of input.intakeEvents) {
-    if (i.takenAt) tally(berlinHour(i.takenAt), toBerlinDayKey(i.takenAt));
+    if (i.takenAt) tally(hourInTz(i.takenAt, tz), userDayKey(i.takenAt, tz));
  }

  let doctorPdfCount = 0;
@@ -412,22 +410,26 @@ export function buildExpansionMetricValues(input: {
   * hours derive from `measuredAt` exactly as before.
   */
  measurementHours?: number[];
+  /** v1.30 (DATAINT M4) — day-key timezone; defaults to Berlin (see file docblock). */
+  tz?: string;
 }): ExpansionMetrics {
+  const tz = input.tz ?? DEFAULT_TIMEZONE;
  const counts = countMeasurementsByType(input.measurements);
  const mood = getMoodMetrics(input.moodEntries);
-  // The hidden pass also needs the per-row Berlin hour; derive it once here
-  // (only when the caller supplied day-keys, i.e. on the hot achievements
-  // path) so a single hour pass replaces the per-pass re-derivation.
+  // The hidden pass also needs the per-row hour; derive it once here (only
+  // when the caller supplied day-keys, i.e. on the hot achievements path) so
+  // a single hour pass replaces the per-pass re-derivation.
  const measurementHours =
    input.measurementHours ??
    (input.measurementDayKeys
-      ? input.measurements.map((m) => berlinHour(m.measuredAt))
+      ? input.measurements.map((m) => hourInTz(m.measuredAt, tz))
      : undefined);
  const engagement = getEngagementMetrics({
    measurements: input.measurements,
    moodEntries: input.moodEntries,
    intakeEvents: input.intakeEvents,
    measurementDayKeys: input.measurementDayKeys,
+    tz,
  });
  const hidden = getHiddenMetrics({
    measurements: input.measurements,
@@ -436,6 +438,7 @@ export function buildExpansionMetricValues(input: {
    auditEvents: input.auditEvents,
    measurementDayKeys: input.measurementDayKeys,
    measurementHours,
+    tz,
  });

  return {
@@ -209,11 +209,14 @@ const ECG_ENTRY: CatalogEntry = {
 /**
 * Discoverability audit finding A5 — the alert-style watchdogs (rhythm-
 * event notifications, breathing-disturbance screening, baseline-drift
- * health status) are CORRECT to stay unmounted when there is nothing to
- * flag ("no alert" must not render an alarming empty shell), but nothing
- * anywhere told the user they were running. This is the "single line"
- * fix the audit recommended in place of per-card empty states: an
- * always-shown info row, no present/absent gate, no CTA.
+ * health status, and the labs "what changed since your last panel" card)
+ * are CORRECT to stay unmounted when there is nothing to flag ("no
+ * alert" must not render an alarming empty shell), but nothing anywhere
+ * told the user they were running. This is the "single line" fix the
+ * audit recommended in place of per-card empty states: an always-shown
+ * info row, no present/absent gate, no CTA. The description copy covers
+ * all four watchdogs — extend it (not a new catalog row) if a fifth one
+ * ships.
 */
 const WATCHDOG_ENTRY: CatalogEntry = {
  id: "watchdog",
@@ -642,7 +642,7 @@ export const coachPaths: NonNullable<ZodOpenApiObject["paths"]> = {
      tags: ["Insights"],
      summary: "List the caller's Coach conversations",
      description:
-        'v1.18.0 — cursor-paginated list of the caller\'s Coach conversations for the history rail, most-recent activity first. Metadata only (id, title, timestamps, message count); message bodies are not decrypted here. `limit` defaults to 20, capped at 50; pass the returned `nextCursor` back as `cursor` for the next page (null at the end). Coach-gated (`requireAssistantSurface("coach")`); a disabled surface 403s. Auth via cookie or Bearer; the owner is narrowed from the session.',
+        'v1.18.0 — cursor-paginated list of the caller\'s Coach conversations for the history rail, most-recent activity first. Metadata only (id, title, timestamps, message count); message bodies are not decrypted here. `limit` defaults to 20, capped at 50; pass the returned `nextCursor` back as `cursor` for the next page (null at the end). v1.30.2 — optional `q` narrows the page to conversations whose TITLE contains the text (case-insensitive substring, capped at 200 chars); message bodies are encrypted at rest and are not searched. Coach-gated (`requireAssistantSurface("coach")`); a disabled surface 403s. Auth via cookie or Bearer; the owner is narrowed from the session.',
      parameters: [
        {
          name: "cursor",
@@ -659,6 +659,14 @@ export const coachPaths: NonNullable<ZodOpenApiObject["paths"]> = {
          schema: { type: "integer", minimum: 1, maximum: 50, default: 20 },
          description: "Page size. Defaults to 20, capped at 50.",
        },
+        {
+          name: "q",
+          in: "query",
+          required: false,
+          schema: { type: "string", maxLength: 200 },
+          description:
+            "v1.30.2 — case-insensitive substring filter over the conversation TITLE only (message bodies are encrypted and not searched). Omit for the unfiltered list.",
+        },
      ],
      responses: {
        "200": {
@@ -20,6 +20,18 @@ export const coachKeys = {
   * share one factory-routed key.
   */
  coachConversations: () => ["coachConversations"] as const,
+  /**
+   * v1.30.2 — the full, paginated + searchable rail history
+   * (`useInfiniteQuery` over `GET /api/insights/chat?cursor&q`). Nested under
+   * the `coachConversations` prefix (not a sibling key) so the existing
+   * delete-mutation + post-turn invalidation (`coachConversations()`) still
+   * reaches every cached search variant of this infinite list — TanStack's
+   * default `invalidateQueries` match is prefix-based, so no extra
+   * invalidation call site was needed. `search` is the trimmed query text;
+   * empty string keys the unfiltered ("browse everything") variant.
+   */
+  coachConversationHistory: (search = "") =>
+    ["coachConversations", "history", search.trim()] as const,
  /**
   * v1.18.7 — one decrypted Coach conversation (`GET /api/insights/chat/[id]`).
   * Keyed on the conversation id so two open threads never share a cache slot;