Version: 0.9.23.dev.260416

后端： 1. Memory 管理面 API 落地（“我的记忆”增删改查 + 恢复） - 补齐 List/Get/Create/Update/Delete/Restore 的 handler、请求模型与返回视图 - 注册 `/api/v1/memory/items*` 路由并接入 MemoryHandler - 新增 memory item not found / invalid memory type / invalid memory content 三类管理面错误码 2. Memory Module / Service / Repo 扩展为“可管理 + 可治理”门面 - 新增 NewModuleWithObserve / ObserveDeps，导出 GetItem / CreateItem / UpdateItem / DeleteItem / RestoreItem / RunDedupCleanup / MemoryObserver / MemoryMetrics - 新增手动新增、修改、恢复能力；删除链路切到 SoftDeleteByID；所有管理动作统一事务内写 audit，并桥接向量同步与管理面观测 - 补齐 CreateItemFields / UpdateItemFields、单条 Create、管理侧字段更新、软删/恢复，以及 dedup 扫描/归档所需 repo 能力 - 审计操作补齐 archive / restore 3. Memory 读侧与注入侧观测补齐 - HybridRetrieve 返回 telemetry，统一记录 pinned hit / semantic hit / dedup drop / degraded / RAG fallback，并上报读取命中、去重丢弃、RAG 降级指标 - AgentService 持有 memory observer / metrics；injectMemoryContext 对读取失败、空注入、成功注入补齐结构化日志与注入计数 4. Worker / 决策 / 向量同步链路治理增强 - 召回结果显式携带 fallbackMode；hash 精确命中、rag→mysql 降级、最终动作统一写入决策观测 - 接入 vectorSyncer / observer / metrics；为 job 重试、任务成功/失败、决策分布与 fallback 补齐打点；向量 upsert/delete 统一改走公共 Syncer，并收敛 parseMemoryID 解析逻辑 5. 启动层接入 Memory 观测依赖 - 启动时创建 LoggerObserver + MetricsRegistry，并通过 NewModuleWithObserve 注入 memory 模块前端：无仓库：无
2026-04-16 19:34:32 +08:00
parent a1b2ffedb8
commit fad3aed30a
23 changed files with 2527 additions and 121 deletions
--- a/backend/api/container.go
+++ b/backend/api/container.go
@@ -7,4 +7,5 @@ type ApiHandlers struct {
 	TaskClassHandler *TaskClassHandler
 	ScheduleHandler  *ScheduleAPI
 	AgentHandler     *AgentHandler
+	MemoryHandler    *MemoryHandler
 }
--- a/backend/api/memory.go
+++ b/backend/api/memory.go
@@ -0,0 +1,290 @@
+package api
+
+import (
+	"context"
+	"errors"
+	"net/http"
+	"strconv"
+	"strings"
+	"time"
+
+	memorypkg "github.com/LoveLosita/smartflow/backend/memory"
+	memorymodel "github.com/LoveLosita/smartflow/backend/memory/model"
+	"github.com/LoveLosita/smartflow/backend/model"
+	"github.com/LoveLosita/smartflow/backend/respond"
+	"github.com/gin-gonic/gin"
+)
+
+type MemoryHandler struct {
+	module *memorypkg.Module
+}
+
+var errMemoryHandlerNotReady = errors.New("memory handler is not initialized")
+
+func NewMemoryHandler(module *memorypkg.Module) *MemoryHandler {
+	return &MemoryHandler{module: module}
+}
+
+func (h *MemoryHandler) ListItems(c *gin.Context) {
+	if h == nil || h.module == nil {
+		c.JSON(http.StatusInternalServerError, respond.InternalError(errMemoryHandlerNotReady))
+		return
+	}
+
+	limit, ok := parseOptionalInt(c.Query("limit"))
+	if !ok {
+		c.JSON(http.StatusBadRequest, respond.WrongParamType)
+		return
+	}
+	statusesRaw := c.Query("statuses")
+	if strings.TrimSpace(statusesRaw) == "" {
+		statusesRaw = c.Query("status")
+	}
+	memoryTypesRaw := c.Query("memory_types")
+	if strings.TrimSpace(memoryTypesRaw) == "" {
+		memoryTypesRaw = c.Query("memory_type")
+	}
+
+	ctx, cancel := context.WithTimeout(c.Request.Context(), 2*time.Second)
+	defer cancel()
+
+	items, err := h.module.ListItems(ctx, memorymodel.ListItemsRequest{
+		UserID:         c.GetInt("user_id"),
+		ConversationID: strings.TrimSpace(c.Query("conversation_id")),
+		Statuses:       splitCSV(statusesRaw),
+		MemoryTypes:    splitCSV(memoryTypesRaw),
+		Limit:          limit,
+	})
+	if err != nil {
+		respond.DealWithError(c, err)
+		return
+	}
+
+	c.JSON(http.StatusOK, respond.RespWithData(respond.Ok, toMemoryItemViews(items)))
+}
+
+func (h *MemoryHandler) GetItem(c *gin.Context) {
+	if h == nil || h.module == nil {
+		c.JSON(http.StatusInternalServerError, respond.InternalError(errMemoryHandlerNotReady))
+		return
+	}
+
+	memoryID, ok := parseMemoryIDParam(c)
+	if !ok {
+		c.JSON(http.StatusBadRequest, respond.WrongParamType)
+		return
+	}
+
+	ctx, cancel := context.WithTimeout(c.Request.Context(), 2*time.Second)
+	defer cancel()
+
+	item, err := h.module.GetItem(ctx, model.MemoryGetItemRequest{
+		UserID:   c.GetInt("user_id"),
+		MemoryID: memoryID,
+	})
+	if err != nil {
+		respond.DealWithError(c, err)
+		return
+	}
+	c.JSON(http.StatusOK, respond.RespWithData(respond.Ok, toMemoryItemView(item)))
+}
+
+func (h *MemoryHandler) CreateItem(c *gin.Context) {
+	if h == nil || h.module == nil {
+		c.JSON(http.StatusInternalServerError, respond.InternalError(errMemoryHandlerNotReady))
+		return
+	}
+
+	var req model.MemoryCreateItemRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		c.JSON(http.StatusBadRequest, respond.WrongParamType)
+		return
+	}
+	req.UserID = c.GetInt("user_id")
+	req.OperatorType = "user"
+
+	ctx, cancel := context.WithTimeout(c.Request.Context(), 2*time.Second)
+	defer cancel()
+
+	item, err := h.module.CreateItem(ctx, req)
+	if err != nil {
+		respond.DealWithError(c, err)
+		return
+	}
+	c.JSON(http.StatusOK, respond.RespWithData(respond.Ok, toMemoryItemView(item)))
+}
+
+func (h *MemoryHandler) UpdateItem(c *gin.Context) {
+	if h == nil || h.module == nil {
+		c.JSON(http.StatusInternalServerError, respond.InternalError(errMemoryHandlerNotReady))
+		return
+	}
+
+	memoryID, ok := parseMemoryIDParam(c)
+	if !ok {
+		c.JSON(http.StatusBadRequest, respond.WrongParamType)
+		return
+	}
+
+	var req model.MemoryUpdateItemRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		c.JSON(http.StatusBadRequest, respond.WrongParamType)
+		return
+	}
+	req.UserID = c.GetInt("user_id")
+	req.MemoryID = memoryID
+	req.OperatorType = "user"
+
+	ctx, cancel := context.WithTimeout(c.Request.Context(), 2*time.Second)
+	defer cancel()
+
+	item, err := h.module.UpdateItem(ctx, req)
+	if err != nil {
+		respond.DealWithError(c, err)
+		return
+	}
+	c.JSON(http.StatusOK, respond.RespWithData(respond.Ok, toMemoryItemView(item)))
+}
+
+func (h *MemoryHandler) DeleteItem(c *gin.Context) {
+	if h == nil || h.module == nil {
+		c.JSON(http.StatusInternalServerError, respond.InternalError(errMemoryHandlerNotReady))
+		return
+	}
+
+	memoryID, ok := parseMemoryIDParam(c)
+	if !ok {
+		c.JSON(http.StatusBadRequest, respond.WrongParamType)
+		return
+	}
+
+	var body struct {
+		Reason string `json:"reason"`
+	}
+	_ = c.ShouldBindJSON(&body)
+
+	ctx, cancel := context.WithTimeout(c.Request.Context(), 2*time.Second)
+	defer cancel()
+
+	item, err := h.module.DeleteItem(ctx, model.MemoryDeleteItemRequest{
+		UserID:       c.GetInt("user_id"),
+		MemoryID:     memoryID,
+		Reason:       strings.TrimSpace(body.Reason),
+		OperatorType: "user",
+	})
+	if err != nil {
+		respond.DealWithError(c, err)
+		return
+	}
+	c.JSON(http.StatusOK, respond.RespWithData(respond.Ok, toMemoryItemView(item)))
+}
+
+func (h *MemoryHandler) RestoreItem(c *gin.Context) {
+	if h == nil || h.module == nil {
+		c.JSON(http.StatusInternalServerError, respond.InternalError(errMemoryHandlerNotReady))
+		return
+	}
+
+	memoryID, ok := parseMemoryIDParam(c)
+	if !ok {
+		c.JSON(http.StatusBadRequest, respond.WrongParamType)
+		return
+	}
+
+	var body struct {
+		Reason string `json:"reason"`
+	}
+	_ = c.ShouldBindJSON(&body)
+
+	ctx, cancel := context.WithTimeout(c.Request.Context(), 2*time.Second)
+	defer cancel()
+
+	item, err := h.module.RestoreItem(ctx, model.MemoryRestoreItemRequest{
+		UserID:       c.GetInt("user_id"),
+		MemoryID:     memoryID,
+		Reason:       strings.TrimSpace(body.Reason),
+		OperatorType: "user",
+	})
+	if err != nil {
+		respond.DealWithError(c, err)
+		return
+	}
+	c.JSON(http.StatusOK, respond.RespWithData(respond.Ok, toMemoryItemView(item)))
+}
+
+func parseMemoryIDParam(c *gin.Context) (int64, bool) {
+	raw := strings.TrimSpace(c.Param("id"))
+	if raw == "" {
+		return 0, false
+	}
+	value, err := strconv.ParseInt(raw, 10, 64)
+	if err != nil || value <= 0 {
+		return 0, false
+	}
+	return value, true
+}
+
+func parseOptionalInt(raw string) (int, bool) {
+	raw = strings.TrimSpace(raw)
+	if raw == "" {
+		return 0, true
+	}
+	value, err := strconv.Atoi(raw)
+	if err != nil {
+		return 0, false
+	}
+	return value, true
+}
+
+func splitCSV(raw string) []string {
+	raw = strings.TrimSpace(raw)
+	if raw == "" {
+		return nil
+	}
+	parts := strings.Split(raw, ",")
+	result := make([]string, 0, len(parts))
+	for _, part := range parts {
+		part = strings.TrimSpace(part)
+		if part == "" {
+			continue
+		}
+		result = append(result, part)
+	}
+	return result
+}
+
+func toMemoryItemViews(items []memorymodel.ItemDTO) []model.MemoryItemView {
+	if len(items) == 0 {
+		return nil
+	}
+	result := make([]model.MemoryItemView, 0, len(items))
+	for _, item := range items {
+		result = append(result, toMemoryItemView(&item))
+	}
+	return result
+}
+
+func toMemoryItemView(item *memorymodel.ItemDTO) model.MemoryItemView {
+	if item == nil {
+		return model.MemoryItemView{}
+	}
+	return model.MemoryItemView{
+		ID:               item.ID,
+		UserID:           item.UserID,
+		ConversationID:   item.ConversationID,
+		AssistantID:      item.AssistantID,
+		RunID:            item.RunID,
+		MemoryType:       item.MemoryType,
+		Title:            item.Title,
+		Content:          item.Content,
+		ContentHash:      item.ContentHash,
+		Confidence:       item.Confidence,
+		Importance:       item.Importance,
+		SensitivityLevel: item.SensitivityLevel,
+		IsExplicit:       item.IsExplicit,
+		Status:           item.Status,
+		TTLAt:            item.TTLAt,
+		CreatedAt:        item.CreatedAt,
+		UpdatedAt:        item.UpdatedAt,
+	}
+}
--- a/backend/cmd/start.go
+++ b/backend/cmd/start.go
@@ -14,6 +14,7 @@ import (
 	ragconfig "github.com/LoveLosita/smartflow/backend/infra/rag/config"
 	"github.com/LoveLosita/smartflow/backend/inits"
 	"github.com/LoveLosita/smartflow/backend/memory"
+	memoryobserve "github.com/LoveLosita/smartflow/backend/memory/observe"
 	"github.com/LoveLosita/smartflow/backend/middleware"
 	newagentconv "github.com/LoveLosita/smartflow/backend/newAgent/conv"
 	newagenttools "github.com/LoveLosita/smartflow/backend/newAgent/tools"
@@ -78,7 +79,18 @@ func Start() {
 	// 1. memory 模块对启动层只暴露一个门面。
 	// 2. 后续若接入统一 DI 容器，也优先注入这个门面，而不是继续暴露内部 repo/service。
 	memoryCfg := memory.LoadConfigFromViper()
-	memoryModule := memory.NewModule(db, infrallm.WrapArkClient(aiHub.Worker), ragRuntime, memoryCfg)
+	memoryObserver := memoryobserve.NewLoggerObserver(log.Default())
+	memoryMetrics := memoryobserve.NewMetricsRegistry()
+	memoryModule := memory.NewModuleWithObserve(
+		db,
+		infrallm.WrapArkClient(aiHub.Worker),
+		ragRuntime,
+		memoryCfg,
+		memory.ObserveDeps{
+			Observer: memoryObserver,
+			Metrics:  memoryMetrics,
+		},
+	)

 	// DAO 层初始化。
 	cacheRepo := dao.NewCacheDAO(rdb)
@@ -180,6 +192,7 @@ func Start() {
 	taskClassApi := api.NewTaskClassHandler(taskClassService)
 	scheduleApi := api.NewScheduleAPI(scheduleService)
 	agentApi := api.NewAgentHandler(agentService)
+	memoryApi := api.NewMemoryHandler(memoryModule)
 	handlers := &api.ApiHandlers{
 		UserHandler:      userApi,
 		TaskHandler:      taskApi,
@@ -187,6 +200,7 @@ func Start() {
 		CourseHandler:    courseApi,
 		ScheduleHandler:  scheduleApi,
 		AgentHandler:     agentApi,
+		MemoryHandler:    memoryApi,
 	}

 	r := routers.RegisterRouters(handlers, cacheRepo, userRepo, limiter)
--- a/backend/memory/cleanup/dedup_policy.go
+++ b/backend/memory/cleanup/dedup_policy.go
@@ -0,0 +1,73 @@
+package cleanup
+
+import (
+	"sort"
+	"time"
+
+	"github.com/LoveLosita/smartflow/backend/model"
+)
+
+const dedupRecentTieWindow = 24 * time.Hour
+
+// DedupDecision 描述单个重复组的治理结论。
+type DedupDecision struct {
+	Keep    model.MemoryItem
+	Archive []model.MemoryItem
+}
+
+// DecideDedupGroup 决定一组重复 active 记忆中“保留谁、归档谁”。
+//
+// 步骤化说明：
+// 1. 先按“最近更新时间”判断谁更值得保留，符合治理计划里的“优先保留最近更新”；
+// 2. 若更新时间非常接近，再比较 confidence/importance，避免刚好相差几秒就误保留低质量版本；
+// 3. 最后用主键逆序兜底，保证同组治理结果稳定可复现。
+func DecideDedupGroup(items []model.MemoryItem) DedupDecision {
+	if len(items) == 0 {
+		return DedupDecision{}
+	}
+
+	ordered := make([]model.MemoryItem, len(items))
+	copy(ordered, items)
+	sort.SliceStable(ordered, func(i, j int) bool {
+		return preferDedupKeep(ordered[i], ordered[j])
+	})
+
+	return DedupDecision{
+		Keep:    ordered[0],
+		Archive: ordered[1:],
+	}
+}
+
+func preferDedupKeep(left model.MemoryItem, right model.MemoryItem) bool {
+	leftTime := dedupBaseTime(left)
+	rightTime := dedupBaseTime(right)
+
+	diff := leftTime.Sub(rightTime)
+	if diff < 0 {
+		diff = -diff
+	}
+	if diff > dedupRecentTieWindow {
+		return leftTime.After(rightTime)
+	}
+
+	if left.Confidence != right.Confidence {
+		return left.Confidence > right.Confidence
+	}
+	if left.Importance != right.Importance {
+		return left.Importance > right.Importance
+	}
+	if !leftTime.Equal(rightTime) {
+		return leftTime.After(rightTime)
+	}
+	return left.ID > right.ID
+}
+
+func dedupBaseTime(item model.MemoryItem) time.Time {
+	if item.UpdatedAt != nil {
+		return *item.UpdatedAt
+	}
+	if item.CreatedAt != nil {
+		return *item.CreatedAt
+	}
+	return time.Time{}
+}
--- a/backend/memory/cleanup/dedup_runner.go
+++ b/backend/memory/cleanup/dedup_runner.go
@@ -0,0 +1,257 @@
+package cleanup
+
+import (
+	"context"
+	"errors"
+	"strconv"
+	"strings"
+	"time"
+
+	memoryobserve "github.com/LoveLosita/smartflow/backend/memory/observe"
+	memoryrepo "github.com/LoveLosita/smartflow/backend/memory/repo"
+	memoryutils "github.com/LoveLosita/smartflow/backend/memory/utils"
+	memoryvectorsync "github.com/LoveLosita/smartflow/backend/memory/vectorsync"
+	"github.com/LoveLosita/smartflow/backend/model"
+	"gorm.io/gorm"
+)
+
+// DedupRunner 负责执行一次离线记忆去重治理。
+//
+// 职责边界：
+// 1. 只处理“active + content_hash 非空”的重复组；
+// 2. 只负责 archive + audit + 向量删除桥接，不负责自动定时调度；
+// 3. 支持 dry-run，便于上线初期先观察治理结果再正式落库。
+type DedupRunner struct {
+	db           *gorm.DB
+	itemRepo     *memoryrepo.ItemRepo
+	auditRepo    *memoryrepo.AuditRepo
+	vectorSyncer *memoryvectorsync.Syncer
+	observer     memoryobserve.Observer
+	metrics      memoryobserve.MetricsRecorder
+}
+
+func NewDedupRunner(
+	db *gorm.DB,
+	itemRepo *memoryrepo.ItemRepo,
+	auditRepo *memoryrepo.AuditRepo,
+	vectorSyncer *memoryvectorsync.Syncer,
+	observer memoryobserve.Observer,
+	metrics memoryobserve.MetricsRecorder,
+) *DedupRunner {
+	if observer == nil {
+		observer = memoryobserve.NewNopObserver()
+	}
+	if metrics == nil {
+		metrics = memoryobserve.NewNopMetrics()
+	}
+	return &DedupRunner{
+		db:           db,
+		itemRepo:     itemRepo,
+		auditRepo:    auditRepo,
+		vectorSyncer: vectorSyncer,
+		observer:     observer,
+		metrics:      metrics,
+	}
+}
+
+// Run 执行一次离线去重治理。
+func (r *DedupRunner) Run(ctx context.Context, req model.MemoryDedupCleanupRequest) (model.MemoryDedupCleanupResult, error) {
+	result := model.MemoryDedupCleanupResult{
+		DryRun: req.DryRun,
+	}
+	if r == nil || r.db == nil || r.itemRepo == nil || r.auditRepo == nil {
+		return result, errors.New("memory dedup runner is not initialized")
+	}
+
+	items, err := r.itemRepo.ListActiveItemsForDedup(ctx, req.UserID, req.Limit)
+	if err != nil {
+		r.recordDedupObserve(ctx, req, result, false, err)
+		return result, err
+	}
+
+	groups := groupDuplicateItems(items)
+	result.ScannedGroupCount = len(groups)
+	if len(groups) == 0 {
+		r.recordDedupObserve(ctx, req, result, true, nil)
+		return result, nil
+	}
+
+	for _, group := range groups {
+		decision := DecideDedupGroup(group)
+		if decision.Keep.ID > 0 {
+			result.KeptCount++
+		}
+		if len(decision.Archive) == 0 {
+			continue
+		}
+
+		result.DedupedGroupCount++
+		archiveIDs := collectDedupIDs(decision.Archive)
+		result.ArchivedCount += len(archiveIDs)
+		result.ArchivedIDs = append(result.ArchivedIDs, archiveIDs...)
+		if req.DryRun {
+			continue
+		}
+
+		now := time.Now()
+		txErr := r.db.WithContext(ctx).Transaction(func(tx *gorm.DB) error {
+			itemRepo := r.itemRepo.WithTx(tx)
+			auditRepo := r.auditRepo.WithTx(tx)
+
+			if archiveErr := itemRepo.ArchiveByIDsAt(ctx, archiveIDs, now); archiveErr != nil {
+				return archiveErr
+			}
+
+			for _, item := range decision.Archive {
+				after := item
+				after.Status = model.MemoryItemStatusArchived
+				after.UpdatedAt = &now
+				after.VectorStatus = "pending"
+
+				audit := memoryutils.BuildItemAuditLog(
+					item.ID,
+					item.UserID,
+					memoryutils.AuditOperationArchive,
+					normalizeCleanupOperator(req.OperatorType),
+					normalizeCleanupReason(req.Reason),
+					&item,
+					&after,
+				)
+				if createErr := auditRepo.Create(ctx, audit); createErr != nil {
+					return createErr
+				}
+			}
+			return nil
+		})
+		if txErr != nil {
+			r.recordDedupObserve(ctx, req, result, false, txErr)
+			return result, txErr
+		}
+
+		r.vectorSyncer.Delete(ctx, "", archiveIDs)
+		r.metrics.AddCounter(memoryobserve.MetricCleanupArchivedTotal, int64(len(archiveIDs)), map[string]string{
+			"dry_run": "false",
+		})
+	}
+
+	r.recordDedupObserve(ctx, req, result, true, nil)
+	return result, nil
+}
+
+func groupDuplicateItems(items []model.MemoryItem) [][]model.MemoryItem {
+	if len(items) == 0 {
+		return nil
+	}
+
+	result := make([][]model.MemoryItem, 0)
+	currentGroup := make([]model.MemoryItem, 0, 2)
+	currentKey := ""
+	for _, item := range items {
+		key := dedupGroupKey(item)
+		if key == "" {
+			continue
+		}
+		if currentKey == "" || currentKey != key {
+			if len(currentGroup) > 1 {
+				copied := make([]model.MemoryItem, len(currentGroup))
+				copy(copied, currentGroup)
+				result = append(result, copied)
+			}
+			currentKey = key
+			currentGroup = currentGroup[:0]
+		}
+		currentGroup = append(currentGroup, item)
+	}
+	if len(currentGroup) > 1 {
+		copied := make([]model.MemoryItem, len(currentGroup))
+		copy(copied, currentGroup)
+		result = append(result, copied)
+	}
+	return result
+}
+
+func dedupGroupKey(item model.MemoryItem) string {
+	contentHash := strings.TrimSpace(derefString(item.ContentHash))
+	if item.UserID <= 0 || strings.TrimSpace(item.MemoryType) == "" || contentHash == "" {
+		return ""
+	}
+	return strings.Join([]string{
+		strconv.Itoa(item.UserID),
+		item.MemoryType,
+		contentHash,
+	}, "::")
+}
+
+func collectDedupIDs(items []model.MemoryItem) []int64 {
+	ids := make([]int64, 0, len(items))
+	for _, item := range items {
+		if item.ID <= 0 {
+			continue
+		}
+		ids = append(ids, item.ID)
+	}
+	return ids
+}
+
+func normalizeCleanupOperator(operatorType string) string {
+	operatorType = strings.TrimSpace(operatorType)
+	if operatorType == "" {
+		return "system"
+	}
+	return memoryutils.NormalizeOperatorType(operatorType)
+}
+
+func normalizeCleanupReason(reason string) string {
+	reason = strings.TrimSpace(reason)
+	if reason == "" {
+		return "离线 dedup 治理归档重复记忆"
+	}
+	return reason
+}
+
+func derefString(value *string) string {
+	if value == nil {
+		return ""
+	}
+	return strings.TrimSpace(*value)
+}
+
+func (r *DedupRunner) recordDedupObserve(
+	ctx context.Context,
+	req model.MemoryDedupCleanupRequest,
+	result model.MemoryDedupCleanupResult,
+	success bool,
+	err error,
+) {
+	if r == nil {
+		return
+	}
+
+	status := "success"
+	level := memoryobserve.LevelInfo
+	if !success || err != nil {
+		status = "error"
+		level = memoryobserve.LevelWarn
+	}
+
+	r.observer.Observe(ctx, memoryobserve.Event{
+		Level:     level,
+		Component: memoryobserve.ComponentCleanup,
+		Operation: memoryobserve.OperationDedup,
+		Fields: map[string]any{
+			"user_id":             req.UserID,
+			"limit":               req.Limit,
+			"dry_run":             req.DryRun,
+			"scanned_group_count": result.ScannedGroupCount,
+			"deduped_group_count": result.DedupedGroupCount,
+			"archived_count":      result.ArchivedCount,
+			"success":             success && err == nil,
+			"error":               err,
+			"error_code":          memoryobserve.ClassifyError(err),
+		},
+	})
+	r.metrics.AddCounter(memoryobserve.MetricCleanupRunTotal, 1, map[string]string{
+		"operation": "dedup",
+		"status":    status,
+	})
+}
--- a/backend/memory/model/item.go
+++ b/backend/memory/model/item.go
@@ -67,10 +67,39 @@ type ListItemsRequest struct {
 	Limit          int
 }

-// DeleteItemRequest 描述软删除一条记忆时所需的最小参数。
-type DeleteItemRequest struct {
+// CreateItemFields 是 repo 层落库时真正需要的字段集合。
+type CreateItemFields struct {
 	UserID            int
-	MemoryID     int64
-	Reason       string
-	OperatorType string
+	ConversationID    string
+	AssistantID       string
+	RunID             string
+	MemoryType        string
+	Title             string
+	Content           string
+	NormalizedContent string
+	ContentHash       string
+	Confidence        float64
+	Importance        float64
+	SensitivityLevel  int
+	IsExplicit        bool
+	Status            string
+	TTLAt             *time.Time
+	VectorStatus      string
+	SourceMessageID   *int64
+	SourceEventID     *string
+	LastAccessAt      *time.Time
+}
+
+// UpdateItemFields 是“用户管理侧修改记忆”时 repo 层允许更新的字段集合。
+type UpdateItemFields struct {
+	MemoryType        string
+	Title             string
+	Content           string
+	NormalizedContent string
+	ContentHash       string
+	Confidence        float64
+	Importance        float64
+	SensitivityLevel  int
+	IsExplicit        bool
+	TTLAt             *time.Time
 }
--- a/backend/memory/module.go
+++ b/backend/memory/module.go
@@ -7,11 +7,15 @@ import (

 	infrallm "github.com/LoveLosita/smartflow/backend/infra/llm"
 	infrarag "github.com/LoveLosita/smartflow/backend/infra/rag"
+	memorycleanup "github.com/LoveLosita/smartflow/backend/memory/cleanup"
 	memorymodel "github.com/LoveLosita/smartflow/backend/memory/model"
+	memoryobserve "github.com/LoveLosita/smartflow/backend/memory/observe"
 	memoryorchestrator "github.com/LoveLosita/smartflow/backend/memory/orchestrator"
 	memoryrepo "github.com/LoveLosita/smartflow/backend/memory/repo"
 	memoryservice "github.com/LoveLosita/smartflow/backend/memory/service"
+	memoryvectorsync "github.com/LoveLosita/smartflow/backend/memory/vectorsync"
 	memoryworker "github.com/LoveLosita/smartflow/backend/memory/worker"
+	"github.com/LoveLosita/smartflow/backend/model"
 	"gorm.io/gorm"
 )

@@ -26,6 +30,8 @@ type Module struct {
 	cfg        memorymodel.Config
 	llmClient  *infrallm.Client
 	ragRuntime infrarag.Runtime
+	observer   memoryobserve.Observer
+	metrics    memoryobserve.MetricsRecorder

 	jobRepo      *memoryrepo.JobRepo
 	itemRepo     *memoryrepo.ItemRepo
@@ -35,9 +41,17 @@ type Module struct {
 	enqueueService *memoryservice.EnqueueService
 	readService    *memoryservice.ReadService
 	manageService  *memoryservice.ManageService
+	vectorSyncer   *memoryvectorsync.Syncer
+	dedupRunner    *memorycleanup.DedupRunner
 	runner         *memoryworker.Runner
 }

+// ObserveDeps 描述 memory 模块可选的观测依赖。
+type ObserveDeps struct {
+	Observer memoryobserve.Observer
+	Metrics  memoryobserve.MetricsRecorder
+}
+
 // LoadConfigFromViper 复用 memory 子包里的配置加载逻辑，对外收口一个统一入口。
 func LoadConfigFromViper() memorymodel.Config {
 	return memoryservice.LoadConfigFromViper()
@@ -51,7 +65,18 @@ func LoadConfigFromViper() memorymodel.Config {
 // 3. ragRuntime 允许为 nil，此时读取/向量同步自动回退旧逻辑；
 // 4. 若后续接入统一 DI 容器，也应优先注册这个 Module，而不是把内部 repo/service 继续向外泄漏。
 func NewModule(db *gorm.DB, llmClient *infrallm.Client, ragRuntime infrarag.Runtime, cfg memorymodel.Config) *Module {
-	return wireModule(db, llmClient, ragRuntime, cfg)
+	return NewModuleWithObserve(db, llmClient, ragRuntime, cfg, ObserveDeps{})
+}
+
+// NewModuleWithObserve 创建带观测依赖的 memory 模块门面。
+func NewModuleWithObserve(
+	db *gorm.DB,
+	llmClient *infrallm.Client,
+	ragRuntime infrarag.Runtime,
+	cfg memorymodel.Config,
+	deps ObserveDeps,
+) *Module {
+	return wireModule(db, llmClient, ragRuntime, cfg, deps)
 }

 // WithTx 返回绑定到指定事务连接的同构门面。
@@ -67,7 +92,10 @@ func (m *Module) WithTx(tx *gorm.DB) *Module {
 	if tx == nil {
 		return m
 	}
-	return wireModule(tx, m.llmClient, m.ragRuntime, m.cfg)
+	return wireModule(tx, m.llmClient, m.ragRuntime, m.cfg, ObserveDeps{
+		Observer: m.observer,
+		Metrics:  m.metrics,
+	})
 }

 // EnqueueExtract 把一次记忆抽取请求入队到 memory_jobs。
@@ -98,14 +126,46 @@ func (m *Module) ListItems(ctx context.Context, req memorymodel.ListItemsRequest
 	return m.manageService.ListItems(ctx, req)
 }

+// GetItem 返回当前用户自己的单条记忆详情。
+func (m *Module) GetItem(ctx context.Context, req model.MemoryGetItemRequest) (*memorymodel.ItemDTO, error) {
+	if m == nil || m.manageService == nil {
+		return nil, errors.New("memory module manage service is nil")
+	}
+	return m.manageService.GetItem(ctx, req)
+}
+
+// CreateItem 手动新增一条用户记忆。
+func (m *Module) CreateItem(ctx context.Context, req model.MemoryCreateItemRequest) (*memorymodel.ItemDTO, error) {
+	if m == nil || m.manageService == nil {
+		return nil, errors.New("memory module manage service is nil")
+	}
+	return m.manageService.CreateItem(ctx, req)
+}
+
+// UpdateItem 手动修改一条用户记忆。
+func (m *Module) UpdateItem(ctx context.Context, req model.MemoryUpdateItemRequest) (*memorymodel.ItemDTO, error) {
+	if m == nil || m.manageService == nil {
+		return nil, errors.New("memory module manage service is nil")
+	}
+	return m.manageService.UpdateItem(ctx, req)
+}
+
 // DeleteItem 软删除一条记忆，并补写审计日志。
-func (m *Module) DeleteItem(ctx context.Context, req memorymodel.DeleteItemRequest) (*memorymodel.ItemDTO, error) {
+func (m *Module) DeleteItem(ctx context.Context, req model.MemoryDeleteItemRequest) (*memorymodel.ItemDTO, error) {
 	if m == nil || m.manageService == nil {
 		return nil, errors.New("memory module manage service is nil")
 	}
 	return m.manageService.DeleteItem(ctx, req)
 }

+// RestoreItem 恢复一条 deleted/archived 记忆。
+func (m *Module) RestoreItem(ctx context.Context, req model.MemoryRestoreItemRequest) (*memorymodel.ItemDTO, error) {
+	if m == nil || m.manageService == nil {
+		return nil, errors.New("memory module manage service is nil")
+	}
+	return m.manageService.RestoreItem(ctx, req)
+}
+
 // GetUserSetting 读取用户当前生效的记忆开关。
 func (m *Module) GetUserSetting(ctx context.Context, userID int) (memorymodel.UserSettingDTO, error) {
 	if m == nil || m.manageService == nil {
@@ -122,6 +182,30 @@ func (m *Module) UpsertUserSetting(ctx context.Context, req memorymodel.UpdateUs
 	return m.manageService.UpsertUserSetting(ctx, req)
 }

+// RunDedupCleanup 执行一次离线 dedup 治理。
+func (m *Module) RunDedupCleanup(ctx context.Context, req model.MemoryDedupCleanupRequest) (model.MemoryDedupCleanupResult, error) {
+	if m == nil || m.dedupRunner == nil {
+		return model.MemoryDedupCleanupResult{}, errors.New("memory module dedup runner is nil")
+	}
+	return m.dedupRunner.Run(ctx, req)
+}
+
+// MemoryObserver 暴露 memory 模块当前使用的 observer，供注入桥接等外围能力复用。
+func (m *Module) MemoryObserver() memoryobserve.Observer {
+	if m == nil || m.observer == nil {
+		return memoryobserve.NewNopObserver()
+	}
+	return m.observer
+}
+
+// MemoryMetrics 暴露 memory 模块当前使用的轻量计数器。
+func (m *Module) MemoryMetrics() memoryobserve.MetricsRecorder {
+	if m == nil || m.metrics == nil {
+		return memoryobserve.NewNopMetrics()
+	}
+	return m.metrics
+}
+
 // StartWorker 启动 memory 后台 worker。
 //
 // 说明：
@@ -142,15 +226,30 @@ func (m *Module) StartWorker(ctx context.Context) {
 	log.Println("Memory worker started")
 }

-func wireModule(db *gorm.DB, llmClient *infrallm.Client, ragRuntime infrarag.Runtime, cfg memorymodel.Config) *Module {
+func wireModule(
+	db *gorm.DB,
+	llmClient *infrallm.Client,
+	ragRuntime infrarag.Runtime,
+	cfg memorymodel.Config,
+	deps ObserveDeps,
+) *Module {
 	jobRepo := memoryrepo.NewJobRepo(db)
 	itemRepo := memoryrepo.NewItemRepo(db)
 	auditRepo := memoryrepo.NewAuditRepo(db)
 	settingsRepo := memoryrepo.NewSettingsRepo(db)
+	observer := deps.Observer
+	if observer == nil {
+		observer = memoryobserve.NewLoggerObserver(log.Default())
+	}
+	metrics := deps.Metrics
+	if metrics == nil {
+		metrics = memoryobserve.NewMetricsRegistry()
+	}
+	vectorSyncer := memoryvectorsync.NewSyncer(ragRuntime, itemRepo, observer, metrics)

 	enqueueService := memoryservice.NewEnqueueService(jobRepo)
-	readService := memoryservice.NewReadService(itemRepo, settingsRepo, ragRuntime, cfg)
-	manageService := memoryservice.NewManageService(db, itemRepo, auditRepo, settingsRepo)
+	readService := memoryservice.NewReadService(itemRepo, settingsRepo, ragRuntime, cfg, observer, metrics)
+	manageService := memoryservice.NewManageService(db, itemRepo, auditRepo, settingsRepo, vectorSyncer, observer, metrics)
 	extractor := memoryorchestrator.NewLLMWriteOrchestrator(llmClient, cfg)

 	// 决策编排器：仅在 DecisionEnabled 时才创建有效实例。
@@ -161,13 +260,16 @@ func wireModule(db *gorm.DB, llmClient *infrallm.Client, ragRuntime infrarag.Run
 		decisionOrchestrator = memoryorchestrator.NewLLMDecisionOrchestrator(llmClient, cfg)
 	}

-	runner := memoryworker.NewRunner(db, jobRepo, itemRepo, auditRepo, settingsRepo, extractor, ragRuntime, cfg, decisionOrchestrator)
+	runner := memoryworker.NewRunner(db, jobRepo, itemRepo, auditRepo, settingsRepo, extractor, ragRuntime, cfg, decisionOrchestrator, vectorSyncer, observer, metrics)
+	dedupRunner := memorycleanup.NewDedupRunner(db, itemRepo, auditRepo, vectorSyncer, observer, metrics)

 	return &Module{
 		db:             db,
 		cfg:            cfg,
 		llmClient:      llmClient,
 		ragRuntime:     ragRuntime,
+		observer:       observer,
+		metrics:        metrics,
 		jobRepo:        jobRepo,
 		itemRepo:       itemRepo,
 		auditRepo:      auditRepo,
@@ -175,6 +277,8 @@ func wireModule(db *gorm.DB, llmClient *infrallm.Client, ragRuntime infrarag.Run
 		enqueueService: enqueueService,
 		readService:    readService,
 		manageService:  manageService,
+		vectorSyncer:   vectorSyncer,
+		dedupRunner:    dedupRunner,
 		runner:         runner,
 	}
 }
--- a/backend/memory/observe/log_fields.go
+++ b/backend/memory/observe/log_fields.go
@@ -0,0 +1,119 @@
+package observe
+
+import (
+	"context"
+	"errors"
+	"strings"
+)
+
+const (
+	ComponentRead    = "read"
+	ComponentWrite   = "write"
+	ComponentInject  = "inject"
+	ComponentManage  = "manage"
+	ComponentCleanup = "cleanup"
+
+	OperationRetrieve = "retrieve"
+	OperationDecision = "decision"
+	OperationInject   = "inject"
+	OperationManage   = "manage"
+	OperationDedup    = "dedup"
+
+	MetricJobTotal               = "memory_job_total"
+	MetricJobRetryTotal          = "memory_job_retry_total"
+	MetricDecisionTotal          = "memory_decision_total"
+	MetricDecisionFallbackTotal  = "memory_decision_fallback_total"
+	MetricRetrieveHitTotal       = "memory_retrieve_hit_total"
+	MetricRetrieveDedupDropTotal = "memory_retrieve_dedup_drop_total"
+	MetricInjectItemTotal        = "memory_inject_item_total"
+	MetricRAGFallbackTotal       = "memory_rag_fallback_total"
+	MetricManageTotal            = "memory_manage_total"
+	MetricCleanupRunTotal        = "memory_cleanup_run_total"
+	MetricCleanupArchivedTotal   = "memory_cleanup_archived_total"
+)
+
+type fieldsContextKey struct{}
+
+// WithFields 把 memory 链路公共字段挂进上下文，供下游日志复用。
+//
+// 职责边界：
+// 1. 只负责字段透传与覆盖，不负责真正打印日志；
+// 2. 只保留有意义的字段，避免结构化日志长期堆积空值；
+// 3. 若上游已写入同名字段，则以后写值为准，方便链路逐层补齐上下文。
+func WithFields(ctx context.Context, fields map[string]any) context.Context {
+	if len(fields) == 0 {
+		return ctx
+	}
+	if ctx == nil {
+		ctx = context.Background()
+	}
+
+	merged := FieldsFromContext(ctx)
+	for key, value := range fields {
+		key = strings.TrimSpace(key)
+		if key == "" || !shouldKeepField(value) {
+			continue
+		}
+		merged[key] = value
+	}
+	if len(merged) == 0 {
+		return ctx
+	}
+	return context.WithValue(ctx, fieldsContextKey{}, merged)
+}
+
+// FieldsFromContext 读取当前上下文中已经累积的观测字段。
+func FieldsFromContext(ctx context.Context) map[string]any {
+	if ctx == nil {
+		return map[string]any{}
+	}
+	raw, ok := ctx.Value(fieldsContextKey{}).(map[string]any)
+	if !ok || len(raw) == 0 {
+		return map[string]any{}
+	}
+
+	result := make(map[string]any, len(raw))
+	for key, value := range raw {
+		result[key] = value
+	}
+	return result
+}
+
+// MergeFields 合并多份结构化字段，后写同名字段覆盖先写字段。
+func MergeFields(parts ...map[string]any) map[string]any {
+	result := make(map[string]any)
+	for _, part := range parts {
+		for key, value := range part {
+			key = strings.TrimSpace(key)
+			if key == "" || !shouldKeepField(value) {
+				continue
+			}
+			result[key] = value
+		}
+	}
+	return result
+}
+
+// ClassifyError 把常见错误压成稳定错误码，便于日志与指标统一聚合。
+func ClassifyError(err error) string {
+	switch {
+	case err == nil:
+		return ""
+	case errors.Is(err, context.DeadlineExceeded):
+		return "deadline_exceeded"
+	case errors.Is(err, context.Canceled):
+		return "canceled"
+	default:
+		return "memory_error"
+	}
+}
+
+func shouldKeepField(value any) bool {
+	if value == nil {
+		return false
+	}
+	if text, ok := value.(string); ok {
+		return strings.TrimSpace(text) != ""
+	}
+	return true
+}
--- a/backend/memory/observe/metrics.go
+++ b/backend/memory/observe/metrics.go
@@ -0,0 +1,158 @@
+package observe
+
+import (
+	"sort"
+	"strings"
+	"sync"
+)
+
+// CounterSnapshot 是轻量计数器的快照视图，供后续排障或接平台时读取。
+type CounterSnapshot struct {
+	Name   string
+	Labels map[string]string
+	Value  int64
+}
+
+// MetricsRecorder 描述 memory 模块对计数器的最小依赖。
+type MetricsRecorder interface {
+	AddCounter(name string, delta int64, labels map[string]string)
+	Snapshot() []CounterSnapshot
+}
+
+// NewNopMetrics 返回空实现，保证无观测平台时仍可安全运行。
+func NewNopMetrics() MetricsRecorder {
+	return nopMetrics{}
+}
+
+type nopMetrics struct{}
+
+func (nopMetrics) AddCounter(string, int64, map[string]string) {}
+
+func (nopMetrics) Snapshot() []CounterSnapshot {
+	return nil
+}
+
+// MetricsRegistry 是 memory 模块当前阶段的轻量内存计数器实现。
+//
+// 职责边界：
+// 1. 只做线程安全计数，不负责导出协议；
+// 2. 标签做低基数归一化，避免治理期临时字段把指标打爆；
+// 3. 后续若项目统一接 Prometheus，可直接保留调用口径并替换实现。
+type MetricsRegistry struct {
+	mu       sync.RWMutex
+	counters map[string]*counterRecord
+}
+
+type counterRecord struct {
+	name   string
+	labels map[string]string
+	value  int64
+}
+
+func NewMetricsRegistry() *MetricsRegistry {
+	return &MetricsRegistry{
+		counters: make(map[string]*counterRecord),
+	}
+}
+
+// AddCounter 追加计数值；delta<=0 时直接忽略，避免脏数据污染快照。
+func (r *MetricsRegistry) AddCounter(name string, delta int64, labels map[string]string) {
+	if r == nil || delta <= 0 {
+		return
+	}
+
+	name = strings.TrimSpace(name)
+	if name == "" {
+		return
+	}
+	normalizedLabels := normalizeLabels(labels)
+	key := buildCounterKey(name, normalizedLabels)
+
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	if existing, ok := r.counters[key]; ok {
+		existing.value += delta
+		return
+	}
+	r.counters[key] = &counterRecord{
+		name:   name,
+		labels: normalizedLabels,
+		value:  delta,
+	}
+}
+
+// Snapshot 返回当前全部计数器快照，便于后续排障或测试读取。
+func (r *MetricsRegistry) Snapshot() []CounterSnapshot {
+	if r == nil {
+		return nil
+	}
+
+	r.mu.RLock()
+	defer r.mu.RUnlock()
+	if len(r.counters) == 0 {
+		return nil
+	}
+
+	keys := make([]string, 0, len(r.counters))
+	for key := range r.counters {
+		keys = append(keys, key)
+	}
+	sort.Strings(keys)
+
+	result := make([]CounterSnapshot, 0, len(keys))
+	for _, key := range keys {
+		record := r.counters[key]
+		labels := make(map[string]string, len(record.labels))
+		for labelKey, labelValue := range record.labels {
+			labels[labelKey] = labelValue
+		}
+		result = append(result, CounterSnapshot{
+			Name:   record.name,
+			Labels: labels,
+			Value:  record.value,
+		})
+	}
+	return result
+}
+
+func normalizeLabels(labels map[string]string) map[string]string {
+	if len(labels) == 0 {
+		return nil
+	}
+
+	result := make(map[string]string, len(labels))
+	for key, value := range labels {
+		key = strings.TrimSpace(key)
+		value = strings.TrimSpace(value)
+		if key == "" || value == "" {
+			continue
+		}
+		result[key] = value
+	}
+	if len(result) == 0 {
+		return nil
+	}
+	return result
+}
+
+func buildCounterKey(name string, labels map[string]string) string {
+	if len(labels) == 0 {
+		return name
+	}
+
+	keys := make([]string, 0, len(labels))
+	for key := range labels {
+		keys = append(keys, key)
+	}
+	sort.Strings(keys)
+
+	var sb strings.Builder
+	sb.WriteString(name)
+	for _, key := range keys {
+		sb.WriteString("|")
+		sb.WriteString(key)
+		sb.WriteString("=")
+		sb.WriteString(labels[key])
+	}
+	return sb.String()
+}
--- a/backend/memory/observe/observer.go
+++ b/backend/memory/observe/observer.go
@@ -0,0 +1,109 @@
+package observe
+
+import (
+	"context"
+	"fmt"
+	"log"
+	"sort"
+	"strings"
+)
+
+// Level 表示 memory 结构化观测事件等级。
+type Level string
+
+const (
+	LevelInfo  Level = "info"
+	LevelWarn  Level = "warn"
+	LevelError Level = "error"
+)
+
+// Event 描述一次 memory 模块内部结构化观测事件。
+//
+// 职责边界：
+// 1. 只承载稳定字段，不绑定具体日志平台；
+// 2. 组件与操作名尽量保持低基数，避免后续指标聚合失控；
+// 3. 字段内容应偏“排障与治理”，不承载大段原始文本。
+type Event struct {
+	Level     Level
+	Component string
+	Operation string
+	Fields    map[string]any
+}
+
+// Observer 是 memory 模块的最小观测接口。
+type Observer interface {
+	Observe(ctx context.Context, event Event)
+}
+
+// ObserverFunc 允许用函数快速适配 Observer。
+type ObserverFunc func(ctx context.Context, event Event)
+
+func (f ObserverFunc) Observe(ctx context.Context, event Event) {
+	if f == nil {
+		return
+	}
+	f(ctx, event)
+}
+
+// NewNopObserver 返回空实现，保证观测能力不会反向阻塞主链路。
+func NewNopObserver() Observer {
+	return ObserverFunc(func(context.Context, Event) {})
+}
+
+// NewLoggerObserver 返回标准日志实现，当前阶段默认打到后端进程日志。
+func NewLoggerObserver(logger *log.Logger) Observer {
+	if logger == nil {
+		logger = log.Default()
+	}
+	return &loggerObserver{logger: logger}
+}
+
+type loggerObserver struct {
+	logger *log.Logger
+}
+
+func (o *loggerObserver) Observe(ctx context.Context, event Event) {
+	if o == nil || o.logger == nil {
+		return
+	}
+
+	level := strings.TrimSpace(string(event.Level))
+	if level == "" {
+		level = string(LevelInfo)
+	}
+	component := strings.TrimSpace(event.Component)
+	if component == "" {
+		component = "unknown"
+	}
+	operation := strings.TrimSpace(event.Operation)
+	if operation == "" {
+		operation = "unknown"
+	}
+
+	fields := FieldsFromContext(ctx)
+	for key, value := range event.Fields {
+		key = strings.TrimSpace(key)
+		if key == "" || !shouldKeepField(value) {
+			continue
+		}
+		fields[key] = value
+	}
+
+	parts := []string{
+		"memory",
+		fmt.Sprintf("level=%s", level),
+		fmt.Sprintf("component=%s", component),
+		fmt.Sprintf("operation=%s", operation),
+	}
+
+	keys := make([]string, 0, len(fields))
+	for key := range fields {
+		keys = append(keys, key)
+	}
+	sort.Strings(keys)
+	for _, key := range keys {
+		parts = append(parts, fmt.Sprintf("%s=%v", key, fields[key]))
+	}
+
+	o.logger.Print(strings.Join(parts, " "))
+}
--- a/backend/memory/repo/item_repo.go
+++ b/backend/memory/repo/item_repo.go
@@ -46,6 +46,54 @@ func (r *ItemRepo) UpsertItems(ctx context.Context, items []model.MemoryItem) er
 	return nil
 }

+// Create 写入单条记忆并返回带自增主键的结果。
+//
+// 职责边界：
+// 1. 只负责单条落库，不负责内容归一化与业务校验；
+// 2. 默认把 vector_status 视为上游已决策好的桥接状态，不在这里擅自改写；
+// 3. 返回值用于上游继续写 audit 或做向量同步。
+func (r *ItemRepo) Create(ctx context.Context, fields memorymodel.CreateItemFields) (*model.MemoryItem, error) {
+	if r == nil || r.db == nil {
+		return nil, errors.New("memory item repo is nil")
+	}
+	if fields.UserID <= 0 {
+		return nil, errors.New("memory item create user_id is invalid")
+	}
+
+	item := model.MemoryItem{
+		UserID:            fields.UserID,
+		ConversationID:    strPtrOrNil(fields.ConversationID),
+		AssistantID:       strPtrOrNil(fields.AssistantID),
+		RunID:             strPtrOrNil(fields.RunID),
+		MemoryType:        fields.MemoryType,
+		Title:             fields.Title,
+		Content:           fields.Content,
+		NormalizedContent: strPtrOrNil(fields.NormalizedContent),
+		ContentHash:       strPtrOrNil(fields.ContentHash),
+		Confidence:        fields.Confidence,
+		Importance:        fields.Importance,
+		SensitivityLevel:  fields.SensitivityLevel,
+		SourceMessageID:   fields.SourceMessageID,
+		SourceEventID:     fields.SourceEventID,
+		IsExplicit:        fields.IsExplicit,
+		Status:            fields.Status,
+		TTLAt:             fields.TTLAt,
+		LastAccessAt:      fields.LastAccessAt,
+		VectorStatus:      fields.VectorStatus,
+	}
+	if item.Status == "" {
+		item.Status = model.MemoryItemStatusActive
+	}
+	if strings.TrimSpace(item.VectorStatus) == "" {
+		item.VectorStatus = "pending"
+	}
+
+	if err := r.db.WithContext(ctx).Create(&item).Error; err != nil {
+		return nil, err
+	}
+	return &item, nil
+}
+
 // FindByQuery 按统一过滤条件读取记忆条目。
 //
 // 步骤化说明：
@@ -324,6 +372,53 @@ func (r *ItemRepo) UpdateContentByID(ctx context.Context, memoryID int64, fields
 		}).Error
 }

+// UpdateManagedFieldsByID 更新“用户管理侧”允许修改的记忆字段。
+func (r *ItemRepo) UpdateManagedFieldsByID(ctx context.Context, userID int, memoryID int64, fields memorymodel.UpdateItemFields) error {
+	return r.UpdateManagedFieldsByIDAt(ctx, userID, memoryID, fields, time.Now())
+}
+
+// UpdateManagedFieldsByIDAt 更新“用户管理侧”允许修改的记忆字段，并允许显式指定更新时间。
+//
+// 步骤化说明：
+// 1. 这里只改内容侧和展示侧字段，不改 user_id/status 等归属语义；
+// 2. memory_type/content 变化后，会把 vector_status 置为 pending，提示上游需要重新同步向量；
+// 3. TTLAt 允许被设置为 nil，用于显式清空过期时间。
+func (r *ItemRepo) UpdateManagedFieldsByIDAt(
+	ctx context.Context,
+	userID int,
+	memoryID int64,
+	fields memorymodel.UpdateItemFields,
+	updatedAt time.Time,
+) error {
+	if r == nil || r.db == nil {
+		return errors.New("memory item repo is nil")
+	}
+	if userID <= 0 || memoryID <= 0 {
+		return errors.New("memory item update params is invalid")
+	}
+	if updatedAt.IsZero() {
+		updatedAt = time.Now()
+	}
+
+	return r.db.WithContext(ctx).
+		Model(&model.MemoryItem{}).
+		Where("id = ? AND user_id = ?", memoryID, userID).
+		Updates(map[string]any{
+			"memory_type":        fields.MemoryType,
+			"title":              fields.Title,
+			"content":            fields.Content,
+			"normalized_content": fields.NormalizedContent,
+			"content_hash":       fields.ContentHash,
+			"confidence":         fields.Confidence,
+			"importance":         fields.Importance,
+			"sensitivity_level":  fields.SensitivityLevel,
+			"is_explicit":        fields.IsExplicit,
+			"ttl_at":             fields.TTLAt,
+			"vector_status":      "pending",
+			"updated_at":         updatedAt,
+		}).Error
+}
+
 // SoftDeleteByID 软删除指定用户的某条记忆。
 //
 // 说明：
@@ -348,6 +443,95 @@ func (r *ItemRepo) SoftDeleteByID(ctx context.Context, userID int, memoryID int6
 		}).Error
 }

+// RestoreByID 把 deleted/archived 记忆恢复为 active。
+func (r *ItemRepo) RestoreByID(ctx context.Context, userID int, memoryID int64) error {
+	return r.RestoreByIDAt(ctx, userID, memoryID, time.Now())
+}
+
+// RestoreByIDAt 把 deleted/archived 记忆恢复为 active，并显式刷新 vector_status。
+//
+// 这样做的原因：
+// 1. 恢复后的记忆需要重新参与语义召回，因此向量侧也要重新同步；
+// 2. 这里统一把 vector_status 置为 pending，避免上游遗漏桥接状态更新；
+// 3. 若目标记录本身已是 active，上游应先读快照决定是否真的调用恢复。
+func (r *ItemRepo) RestoreByIDAt(ctx context.Context, userID int, memoryID int64, updatedAt time.Time) error {
+	if r == nil || r.db == nil {
+		return errors.New("memory item repo is nil")
+	}
+	if userID <= 0 || memoryID <= 0 {
+		return errors.New("memory item restore params is invalid")
+	}
+	if updatedAt.IsZero() {
+		updatedAt = time.Now()
+	}
+
+	return r.db.WithContext(ctx).
+		Model(&model.MemoryItem{}).
+		Where("id = ? AND user_id = ?", memoryID, userID).
+		Updates(map[string]any{
+			"status":        model.MemoryItemStatusActive,
+			"vector_status": "pending",
+			"updated_at":    updatedAt,
+		}).Error
+}
+
+// ArchiveByIDsAt 把一批重复记忆改为 archived，并等待上游删除向量副本。
+func (r *ItemRepo) ArchiveByIDsAt(ctx context.Context, ids []int64, updatedAt time.Time) error {
+	if r == nil || r.db == nil {
+		return errors.New("memory item repo is nil")
+	}
+	if len(ids) == 0 {
+		return nil
+	}
+	if updatedAt.IsZero() {
+		updatedAt = time.Now()
+	}
+
+	return r.db.WithContext(ctx).
+		Model(&model.MemoryItem{}).
+		Where("id IN ?", ids).
+		Where("status = ?", model.MemoryItemStatusActive).
+		Updates(map[string]any{
+			"status":        model.MemoryItemStatusArchived,
+			"vector_status": "pending",
+			"updated_at":    updatedAt,
+		}).Error
+}
+
+// ListActiveItemsForDedup 读取“当前仍 active 且带 content_hash”的候选记忆，供离线 dedup 治理使用。
+//
+// 步骤化说明：
+// 1. 只扫描 status=active 且 hash 非空的记录，因为治理目标是“活跃重复项”；
+// 2. 先按 user/type/hash 分组，再按更新时间、置信度、主键逆序排列，方便上游顺序分组；
+// 3. Limit 仅用于保守控量，不保证整组完整，因此首次治理建议留空或给足够大值。
+func (r *ItemRepo) ListActiveItemsForDedup(ctx context.Context, userID int, limit int) ([]model.MemoryItem, error) {
+	if r == nil || r.db == nil {
+		return nil, errors.New("memory item repo is nil")
+	}
+
+	db := r.db.WithContext(ctx).
+		Model(&model.MemoryItem{}).
+		Where("status = ?", model.MemoryItemStatusActive).
+		Where("content_hash IS NOT NULL AND content_hash <> ''")
+	if userID > 0 {
+		db = db.Where("user_id = ?", userID)
+	}
+	if limit > 0 {
+		db = db.Limit(limit)
+	}
+
+	var items []model.MemoryItem
+	err := db.
+		Order("user_id ASC").
+		Order("memory_type ASC").
+		Order("content_hash ASC").
+		Order("updated_at DESC").
+		Order("confidence DESC").
+		Order("id DESC").
+		Find(&items).Error
+	return items, err
+}
+
 func applyScopedEquality(db *gorm.DB, column, value string, includeGlobal bool) *gorm.DB {
 	value = strings.TrimSpace(value)
 	if value == "" {
--- a/backend/memory/service/manage_service.go
+++ b/backend/memory/service/manage_service.go
@@ -7,15 +7,20 @@ import (
 	"time"

 	memorymodel "github.com/LoveLosita/smartflow/backend/memory/model"
+	memoryobserve "github.com/LoveLosita/smartflow/backend/memory/observe"
 	memoryrepo "github.com/LoveLosita/smartflow/backend/memory/repo"
 	memoryutils "github.com/LoveLosita/smartflow/backend/memory/utils"
+	memoryvectorsync "github.com/LoveLosita/smartflow/backend/memory/vectorsync"
 	"github.com/LoveLosita/smartflow/backend/model"
+	"github.com/LoveLosita/smartflow/backend/respond"
 	"gorm.io/gorm"
 )

 const (
 	defaultManageListLimit  = 20
 	maxManageListLimit      = 100
+	defaultManualConfidence = 0.95
+	defaultManualImportance = 0.90
 )

 // ManageService 负责 memory 模块内部的管理面能力。
@@ -29,6 +34,9 @@ type ManageService struct {
 	itemRepo     *memoryrepo.ItemRepo
 	auditRepo    *memoryrepo.AuditRepo
 	settingsRepo *memoryrepo.SettingsRepo
+	vectorSyncer *memoryvectorsync.Syncer
+	observer     memoryobserve.Observer
+	metrics      memoryobserve.MetricsRecorder
 }

 func NewManageService(
@@ -36,12 +44,24 @@ func NewManageService(
 	itemRepo *memoryrepo.ItemRepo,
 	auditRepo *memoryrepo.AuditRepo,
 	settingsRepo *memoryrepo.SettingsRepo,
+	vectorSyncer *memoryvectorsync.Syncer,
+	observer memoryobserve.Observer,
+	metrics memoryobserve.MetricsRecorder,
 ) *ManageService {
+	if observer == nil {
+		observer = memoryobserve.NewNopObserver()
+	}
+	if metrics == nil {
+		metrics = memoryobserve.NewNopMetrics()
+	}
 	return &ManageService{
 		db:           db,
 		itemRepo:     itemRepo,
 		auditRepo:    auditRepo,
 		settingsRepo: settingsRepo,
+		vectorSyncer: vectorSyncer,
+		observer:     observer,
+		metrics:      metrics,
 	}
 }

@@ -77,18 +97,152 @@ func (s *ManageService) ListItems(ctx context.Context, req memorymodel.ListItems
 	return toItemDTOs(items), nil
 }

+// GetItem 返回“当前用户自己的某条记忆”详情。
+func (s *ManageService) GetItem(ctx context.Context, req model.MemoryGetItemRequest) (*memorymodel.ItemDTO, error) {
+	if s == nil || s.itemRepo == nil {
+		return nil, errors.New("memory manage service is nil")
+	}
+	if req.UserID <= 0 {
+		return nil, respond.WrongUserID
+	}
+	if req.MemoryID <= 0 {
+		return nil, respond.WrongParamType
+	}
+
+	item, err := s.itemRepo.GetByIDForUser(ctx, req.UserID, req.MemoryID)
+	if err != nil {
+		return nil, translateManageError(err)
+	}
+	dto := toItemDTO(*item)
+	return &dto, nil
+}
+
+// CreateItem 手动新增一条用户记忆，并补审计与向量同步桥接。
+func (s *ManageService) CreateItem(ctx context.Context, req model.MemoryCreateItemRequest) (*memorymodel.ItemDTO, error) {
+	if s == nil || s.db == nil || s.itemRepo == nil || s.auditRepo == nil {
+		return nil, errors.New("memory manage service is not initialized")
+	}
+	if req.UserID <= 0 {
+		return nil, respond.WrongUserID
+	}
+
+	fields, err := buildCreateItemFields(req)
+	if err != nil {
+		s.recordManageAction(ctx, "create", req.UserID, 0, fields.MemoryType, false, err)
+		return nil, err
+	}
+
+	var createdItem model.MemoryItem
+	err = s.db.WithContext(ctx).Transaction(func(tx *gorm.DB) error {
+		itemRepo := s.itemRepo.WithTx(tx)
+		auditRepo := s.auditRepo.WithTx(tx)
+
+		created, createErr := itemRepo.Create(ctx, fields)
+		if createErr != nil {
+			return createErr
+		}
+		createdItem = *created
+
+		audit := memoryutils.BuildItemAuditLog(
+			createdItem.ID,
+			createdItem.UserID,
+			memoryutils.AuditOperationCreate,
+			memoryutils.NormalizeOperatorType(req.OperatorType),
+			normalizeManageReason(req.Reason, "用户手动新增记忆"),
+			nil,
+			&createdItem,
+		)
+		return auditRepo.Create(ctx, audit)
+	})
+	if err != nil {
+		err = translateManageError(err)
+		s.recordManageAction(ctx, "create", req.UserID, 0, fields.MemoryType, false, err)
+		return nil, err
+	}
+
+	s.vectorSyncer.Upsert(ctx, "", []model.MemoryItem{createdItem})
+	s.recordManageAction(ctx, "create", req.UserID, createdItem.ID, createdItem.MemoryType, true, nil)
+	dto := toItemDTO(createdItem)
+	return &dto, nil
+}
+
+// UpdateItem 手动修改一条用户记忆，并补审计与向量重同步桥接。
+func (s *ManageService) UpdateItem(ctx context.Context, req model.MemoryUpdateItemRequest) (*memorymodel.ItemDTO, error) {
+	if s == nil || s.db == nil || s.itemRepo == nil || s.auditRepo == nil {
+		return nil, errors.New("memory manage service is not initialized")
+	}
+	if req.UserID <= 0 {
+		return nil, respond.WrongUserID
+	}
+	if req.MemoryID <= 0 {
+		return nil, respond.WrongParamType
+	}
+
+	var updatedItem model.MemoryItem
+	err := s.db.WithContext(ctx).Transaction(func(tx *gorm.DB) error {
+		itemRepo := s.itemRepo.WithTx(tx)
+		auditRepo := s.auditRepo.WithTx(tx)
+
+		current, getErr := itemRepo.GetByIDForUser(ctx, req.UserID, req.MemoryID)
+		if getErr != nil {
+			return getErr
+		}
+
+		fields, afterItem, buildErr := buildUpdateItemFields(req, *current)
+		if buildErr != nil {
+			return buildErr
+		}
+		now := time.Now()
+		afterItem.UpdatedAt = &now
+		afterItem.VectorStatus = "pending"
+
+		if updateErr := itemRepo.UpdateManagedFieldsByIDAt(ctx, req.UserID, req.MemoryID, fields, now); updateErr != nil {
+			return updateErr
+		}
+
+		audit := memoryutils.BuildItemAuditLog(
+			current.ID,
+			current.UserID,
+			memoryutils.AuditOperationUpdate,
+			memoryutils.NormalizeOperatorType(req.OperatorType),
+			normalizeManageReason(req.Reason, "用户手动修改记忆"),
+			current,
+			&afterItem,
+		)
+		if auditErr := auditRepo.Create(ctx, audit); auditErr != nil {
+			return auditErr
+		}
+
+		updatedItem = afterItem
+		return nil
+	})
+	if err != nil {
+		err = translateManageError(err)
+		s.recordManageAction(ctx, "update", req.UserID, req.MemoryID, resolveUpdateMemoryType(req), false, err)
+		return nil, err
+	}
+
+	s.vectorSyncer.Upsert(ctx, "", []model.MemoryItem{updatedItem})
+	s.recordManageAction(ctx, "update", req.UserID, updatedItem.ID, updatedItem.MemoryType, true, nil)
+	dto := toItemDTO(updatedItem)
+	return &dto, nil
+}
+
 // DeleteItem 软删除一条记忆，并补写审计日志。
 //
 // 步骤化说明：
 // 1. 先在事务里读取当前条目快照，确保审计前镜像和实际删除对象一致；
 // 2. 若该条目已是 deleted，则直接按幂等语义返回，避免重复写多条删除审计；
 // 3. 状态更新成功后再写 audit log，保证“有删除就有审计”，失败时整笔事务回滚。
-func (s *ManageService) DeleteItem(ctx context.Context, req memorymodel.DeleteItemRequest) (*memorymodel.ItemDTO, error) {
+func (s *ManageService) DeleteItem(ctx context.Context, req model.MemoryDeleteItemRequest) (*memorymodel.ItemDTO, error) {
 	if s == nil || s.db == nil || s.itemRepo == nil || s.auditRepo == nil {
 		return nil, errors.New("memory manage service is not initialized")
 	}
-	if req.UserID <= 0 || req.MemoryID <= 0 {
-		return nil, nil
+	if req.UserID <= 0 {
+		return nil, respond.WrongUserID
+	}
+	if req.MemoryID <= 0 {
+		return nil, respond.WrongParamType
 	}

 	now := time.Now()
@@ -113,8 +267,9 @@ func (s *ManageService) DeleteItem(ctx context.Context, req memorymodel.DeleteIt
 		after := before
 		after.Status = model.MemoryItemStatusDeleted
 		after.UpdatedAt = &now
+		after.VectorStatus = "pending"

-		if err = itemRepo.UpdateStatusByIDAt(ctx, req.UserID, req.MemoryID, model.MemoryItemStatusDeleted, now); err != nil {
+		if err = itemRepo.SoftDeleteByID(ctx, req.UserID, req.MemoryID); err != nil {
 			return err
 		}

@@ -135,16 +290,87 @@ func (s *ManageService) DeleteItem(ctx context.Context, req memorymodel.DeleteIt
 		return nil
 	})
 	if err != nil {
+		err = translateManageError(err)
+		s.recordManageAction(ctx, "delete", req.UserID, req.MemoryID, "", false, err)
 		return nil, err
 	}
 	if deletedItem.ID <= 0 {
 		return nil, nil
 	}

+	if deletedItem.Status == model.MemoryItemStatusDeleted {
+		s.vectorSyncer.Delete(ctx, "", []int64{deletedItem.ID})
+	}
+	s.recordManageAction(ctx, "delete", req.UserID, deletedItem.ID, deletedItem.MemoryType, true, nil)
 	result := toItemDTO(deletedItem)
 	return &result, nil
 }

+// RestoreItem 把 archived/deleted 记忆恢复为 active，并补审计与向量同步桥接。
+func (s *ManageService) RestoreItem(ctx context.Context, req model.MemoryRestoreItemRequest) (*memorymodel.ItemDTO, error) {
+	if s == nil || s.db == nil || s.itemRepo == nil || s.auditRepo == nil {
+		return nil, errors.New("memory manage service is not initialized")
+	}
+	if req.UserID <= 0 {
+		return nil, respond.WrongUserID
+	}
+	if req.MemoryID <= 0 {
+		return nil, respond.WrongParamType
+	}
+
+	var restoredItem model.MemoryItem
+	err := s.db.WithContext(ctx).Transaction(func(tx *gorm.DB) error {
+		itemRepo := s.itemRepo.WithTx(tx)
+		auditRepo := s.auditRepo.WithTx(tx)
+
+		current, getErr := itemRepo.GetByIDForUser(ctx, req.UserID, req.MemoryID)
+		if getErr != nil {
+			return getErr
+		}
+		if current.Status == model.MemoryItemStatusActive {
+			restoredItem = *current
+			return nil
+		}
+
+		now := time.Now()
+		before := *current
+		after := before
+		after.Status = model.MemoryItemStatusActive
+		after.UpdatedAt = &now
+		after.VectorStatus = "pending"
+
+		if restoreErr := itemRepo.RestoreByIDAt(ctx, req.UserID, req.MemoryID, now); restoreErr != nil {
+			return restoreErr
+		}
+
+		audit := memoryutils.BuildItemAuditLog(
+			before.ID,
+			before.UserID,
+			memoryutils.AuditOperationRestore,
+			memoryutils.NormalizeOperatorType(req.OperatorType),
+			normalizeManageReason(req.Reason, "用户恢复记忆"),
+			&before,
+			&after,
+		)
+		if auditErr := auditRepo.Create(ctx, audit); auditErr != nil {
+			return auditErr
+		}
+
+		restoredItem = after
+		return nil
+	})
+	if err != nil {
+		err = translateManageError(err)
+		s.recordManageAction(ctx, "restore", req.UserID, req.MemoryID, "", false, err)
+		return nil, err
+	}
+
+	s.vectorSyncer.Upsert(ctx, "", []model.MemoryItem{restoredItem})
+	s.recordManageAction(ctx, "restore", req.UserID, restoredItem.ID, restoredItem.MemoryType, true, nil)
+	dto := toItemDTO(restoredItem)
+	return &dto, nil
+}
+
 // GetUserSetting 返回用户当前生效的记忆开关。
 //
 // 返回语义：
@@ -201,3 +427,233 @@ func normalizeDeleteReason(reason string) string {
 	}
 	return reason
 }
+
+func normalizeManageReason(reason string, fallback string) string {
+	reason = strings.TrimSpace(reason)
+	if reason == "" {
+		return fallback
+	}
+	return reason
+}
+
+func translateManageError(err error) error {
+	switch {
+	case err == nil:
+		return nil
+	case errors.Is(err, gorm.ErrRecordNotFound):
+		return respond.MemoryItemNotFound
+	default:
+		return err
+	}
+}
+
+func buildCreateItemFields(req model.MemoryCreateItemRequest) (memorymodel.CreateItemFields, error) {
+	memoryType, err := normalizeManagedMemoryType(req.MemoryType)
+	if err != nil {
+		return memorymodel.CreateItemFields{}, err
+	}
+	content, normalizedContent, err := normalizeManagedContent(req.Content)
+	if err != nil {
+		return memorymodel.CreateItemFields{}, err
+	}
+
+	title := normalizeManagedTitle(req.Title, content)
+	return memorymodel.CreateItemFields{
+		UserID:            req.UserID,
+		ConversationID:    strings.TrimSpace(req.ConversationID),
+		AssistantID:       strings.TrimSpace(req.AssistantID),
+		RunID:             strings.TrimSpace(req.RunID),
+		MemoryType:        memoryType,
+		Title:             title,
+		Content:           content,
+		NormalizedContent: normalizedContent,
+		ContentHash:       memoryutils.HashContent(memoryType, normalizedContent),
+		Confidence:        normalizeManageScore(req.Confidence, defaultManualConfidence),
+		Importance:        normalizeManageScore(req.Importance, defaultManualImportance),
+		SensitivityLevel:  normalizeManageSensitivity(req.SensitivityLevel, 0),
+		IsExplicit:        normalizeManageBool(req.IsExplicit, true),
+		Status:            model.MemoryItemStatusActive,
+		TTLAt:             req.TTLAt,
+		VectorStatus:      "pending",
+	}, nil
+}
+
+func buildUpdateItemFields(
+	req model.MemoryUpdateItemRequest,
+	current model.MemoryItem,
+) (memorymodel.UpdateItemFields, model.MemoryItem, error) {
+	memoryType := current.MemoryType
+	if req.MemoryType != nil {
+		normalizedType, err := normalizeManagedMemoryType(*req.MemoryType)
+		if err != nil {
+			return memorymodel.UpdateItemFields{}, model.MemoryItem{}, err
+		}
+		memoryType = normalizedType
+	}
+
+	content := current.Content
+	if req.Content != nil {
+		normalizedContentValue, _, err := normalizeManagedContent(*req.Content)
+		if err != nil {
+			return memorymodel.UpdateItemFields{}, model.MemoryItem{}, err
+		}
+		content = normalizedContentValue
+	}
+	normalizedContent := normalizeContentForHash(content)
+	if normalizedContent == "" {
+		return memorymodel.UpdateItemFields{}, model.MemoryItem{}, respond.MemoryInvalidContent
+	}
+
+	title := current.Title
+	if req.Title != nil {
+		title = normalizeManagedTitle(*req.Title, content)
+	}
+	ttlAt := current.TTLAt
+	if req.ClearTTL {
+		ttlAt = nil
+	} else if req.TTLAt != nil {
+		ttlAt = req.TTLAt
+	}
+
+	fields := memorymodel.UpdateItemFields{
+		MemoryType:        memoryType,
+		Title:             title,
+		Content:           content,
+		NormalizedContent: normalizedContent,
+		ContentHash:       memoryutils.HashContent(memoryType, normalizedContent),
+		Confidence:        normalizeManageScore(req.Confidence, current.Confidence),
+		Importance:        normalizeManageScore(req.Importance, current.Importance),
+		SensitivityLevel:  normalizeManageSensitivity(req.SensitivityLevel, current.SensitivityLevel),
+		IsExplicit:        normalizeManageBool(req.IsExplicit, current.IsExplicit),
+		TTLAt:             ttlAt,
+	}
+
+	after := current
+	after.MemoryType = fields.MemoryType
+	after.Title = fields.Title
+	after.Content = fields.Content
+	after.NormalizedContent = strPtr(fields.NormalizedContent)
+	after.ContentHash = strPtr(fields.ContentHash)
+	after.Confidence = fields.Confidence
+	after.Importance = fields.Importance
+	after.SensitivityLevel = fields.SensitivityLevel
+	after.IsExplicit = fields.IsExplicit
+	after.TTLAt = fields.TTLAt
+	return fields, after, nil
+}
+
+func normalizeManagedMemoryType(raw string) (string, error) {
+	normalized := memorymodel.NormalizeMemoryType(raw)
+	if normalized == "" {
+		return "", respond.MemoryInvalidType
+	}
+	return normalized, nil
+}
+
+func normalizeManagedContent(raw string) (string, string, error) {
+	content := strings.TrimSpace(raw)
+	if content == "" {
+		return "", "", respond.MemoryInvalidContent
+	}
+	normalized := normalizeContentForHash(content)
+	if normalized == "" {
+		return "", "", respond.MemoryInvalidContent
+	}
+	return content, normalized, nil
+}
+
+func normalizeManagedTitle(raw string, content string) string {
+	title := strings.TrimSpace(raw)
+	if title != "" {
+		return title
+	}
+	content = strings.TrimSpace(content)
+	if content == "" {
+		return "未命名记忆"
+	}
+	runes := []rune(content)
+	if len(runes) > 24 {
+		return string(runes[:24])
+	}
+	return content
+}
+
+func normalizeManageScore(value *float64, defaultValue float64) float64 {
+	if value == nil {
+		return clamp01(defaultValue)
+	}
+	return clamp01(*value)
+}
+
+func normalizeManageSensitivity(value *int, defaultValue int) int {
+	if value == nil {
+		return defaultValue
+	}
+	if *value < 0 {
+		return defaultValue
+	}
+	return *value
+}
+
+func normalizeManageBool(value *bool, defaultValue bool) bool {
+	if value == nil {
+		return defaultValue
+	}
+	return *value
+}
+
+func resolveUpdateMemoryType(req model.MemoryUpdateItemRequest) string {
+	if req.MemoryType == nil {
+		return ""
+	}
+	return strings.TrimSpace(*req.MemoryType)
+}
+
+func strPtr(value string) *string {
+	value = strings.TrimSpace(value)
+	if value == "" {
+		return nil
+	}
+	result := value
+	return &result
+}
+
+func (s *ManageService) recordManageAction(
+	ctx context.Context,
+	operation string,
+	userID int,
+	memoryID int64,
+	memoryType string,
+	success bool,
+	err error,
+) {
+	if s == nil {
+		return
+	}
+
+	status := "success"
+	level := memoryobserve.LevelInfo
+	if !success || err != nil {
+		status = "error"
+		level = memoryobserve.LevelWarn
+	}
+
+	s.metrics.AddCounter(memoryobserve.MetricManageTotal, 1, map[string]string{
+		"operation": strings.TrimSpace(operation),
+		"status":    status,
+	})
+	s.observer.Observe(ctx, memoryobserve.Event{
+		Level:     level,
+		Component: memoryobserve.ComponentManage,
+		Operation: memoryobserve.OperationManage,
+		Fields: map[string]any{
+			"user_id":     userID,
+			"memory_id":   memoryID,
+			"action":      strings.TrimSpace(operation),
+			"memory_type": strings.TrimSpace(memoryType),
+			"success":     success && err == nil,
+			"error":       err,
+			"error_code":  memoryobserve.ClassifyError(err),
+		},
+	})
+}
--- a/backend/memory/service/read_service.go
+++ b/backend/memory/service/read_service.go
@@ -10,6 +10,7 @@ import (

 	infrarag "github.com/LoveLosita/smartflow/backend/infra/rag"
 	memorymodel "github.com/LoveLosita/smartflow/backend/memory/model"
+	memoryobserve "github.com/LoveLosita/smartflow/backend/memory/observe"
 	memoryrepo "github.com/LoveLosita/smartflow/backend/memory/repo"
 	memoryutils "github.com/LoveLosita/smartflow/backend/memory/utils"
 	"github.com/LoveLosita/smartflow/backend/model"
@@ -31,6 +32,26 @@ type ReadService struct {
 	settingsRepo *memoryrepo.SettingsRepo
 	ragRuntime   infrarag.Runtime
 	cfg          memorymodel.Config
+	observer     memoryobserve.Observer
+	metrics      memoryobserve.MetricsRecorder
+}
+
+type retrieveTelemetry struct {
+	ReadMode         string
+	QueryLen         int
+	LegacyHitCount   int
+	PinnedHitCount   int
+	SemanticHitCount int
+	DedupDropCount   int
+	FinalCount       int
+	Degraded         bool
+	RAGFallbackUsed  bool
+}
+
+type semanticRetrieveTelemetry struct {
+	HitCount        int
+	Degraded        bool
+	RAGFallbackUsed bool
 }

 func NewReadService(
@@ -38,12 +59,22 @@ func NewReadService(
 	settingsRepo *memoryrepo.SettingsRepo,
 	ragRuntime infrarag.Runtime,
 	cfg memorymodel.Config,
+	observer memoryobserve.Observer,
+	metrics memoryobserve.MetricsRecorder,
 ) *ReadService {
+	if observer == nil {
+		observer = memoryobserve.NewNopObserver()
+	}
+	if metrics == nil {
+		metrics = memoryobserve.NewNopMetrics()
+	}
 	return &ReadService{
 		itemRepo:     itemRepo,
 		settingsRepo: settingsRepo,
 		ragRuntime:   ragRuntime,
 		cfg:          cfg,
+		observer:     observer,
+		metrics:      metrics,
 	}
 }

@@ -60,9 +91,14 @@ func (s *ReadService) Retrieve(ctx context.Context, req memorymodel.RetrieveRequ
 	if now.IsZero() {
 		now = time.Now()
 	}
+	telemetry := retrieveTelemetry{
+		ReadMode: s.cfg.EffectiveReadMode(),
+		QueryLen: len(strings.TrimSpace(req.Query)),
+	}

 	setting, err := s.settingsRepo.GetByUserID(ctx, req.UserID)
 	if err != nil {
+		s.recordRetrieve(ctx, req, telemetry, err)
 		return nil, err
 	}
 	effectiveSetting := memoryutils.EffectiveUserSetting(setting, req.UserID)
@@ -72,16 +108,29 @@ func (s *ReadService) Retrieve(ctx context.Context, req memorymodel.RetrieveRequ

 	limit := normalizeLimit(req.Limit, defaultRetrieveLimit, maxRetrieveLimit)
 	if s.cfg.EffectiveReadMode() == memorymodel.MemoryReadModeHybrid {
-		return s.HybridRetrieve(ctx, req, effectiveSetting, limit, now)
+		items, hybridTelemetry, hybridErr := s.HybridRetrieve(ctx, req, effectiveSetting, limit, now)
+		hybridTelemetry.ReadMode = memorymodel.MemoryReadModeHybrid
+		hybridTelemetry.QueryLen = telemetry.QueryLen
+		s.recordRetrieve(ctx, req, hybridTelemetry, hybridErr)
+		return items, hybridErr
 	}
 	if s.cfg.RAGEnabled && s.ragRuntime != nil && strings.TrimSpace(req.Query) != "" {
 		items, ragErr := s.retrieveByRAG(ctx, req, effectiveSetting, limit, now)
 		if ragErr == nil && len(items) > 0 {
+			telemetry.SemanticHitCount = len(items)
+			telemetry.FinalCount = len(items)
+			s.recordRetrieve(ctx, req, telemetry, nil)
 			return items, nil
 		}
+		telemetry.Degraded = true
+		telemetry.RAGFallbackUsed = true
 	}

-	return s.retrieveByLegacy(ctx, req, limit, now, effectiveSetting)
+	items, legacyErr := s.retrieveByLegacy(ctx, req, limit, now, effectiveSetting)
+	telemetry.LegacyHitCount = len(items)
+	telemetry.FinalCount = len(items)
+	s.recordRetrieve(ctx, req, telemetry, legacyErr)
+	return items, legacyErr
 }

 func (s *ReadService) retrieveByLegacy(
@@ -180,6 +229,58 @@ func normalizeRetrieveMemoryTypes(raw []string) []string {
 	}
 }

+func (s *ReadService) recordRetrieve(
+	ctx context.Context,
+	req memorymodel.RetrieveRequest,
+	telemetry retrieveTelemetry,
+	err error,
+) {
+	if s == nil {
+		return
+	}
+
+	level := memoryobserve.LevelInfo
+	if err != nil {
+		level = memoryobserve.LevelWarn
+	}
+	s.observer.Observe(ctx, memoryobserve.Event{
+		Level:     level,
+		Component: memoryobserve.ComponentRead,
+		Operation: memoryobserve.OperationRetrieve,
+		Fields: map[string]any{
+			"user_id":            req.UserID,
+			"read_mode":          telemetry.ReadMode,
+			"query_len":          telemetry.QueryLen,
+			"legacy_hit_count":   telemetry.LegacyHitCount,
+			"pinned_hit_count":   telemetry.PinnedHitCount,
+			"semantic_hit_count": telemetry.SemanticHitCount,
+			"dedup_drop_count":   telemetry.DedupDropCount,
+			"final_count":        telemetry.FinalCount,
+			"degraded":           telemetry.Degraded,
+			"rag_fallback_used":  telemetry.RAGFallbackUsed,
+			"success":            err == nil,
+			"error":              err,
+			"error_code":         memoryobserve.ClassifyError(err),
+		},
+	})
+
+	if telemetry.FinalCount > 0 {
+		s.metrics.AddCounter(memoryobserve.MetricRetrieveHitTotal, int64(telemetry.FinalCount), map[string]string{
+			"read_mode": strings.TrimSpace(telemetry.ReadMode),
+		})
+	}
+	if telemetry.DedupDropCount > 0 {
+		s.metrics.AddCounter(memoryobserve.MetricRetrieveDedupDropTotal, int64(telemetry.DedupDropCount), map[string]string{
+			"read_mode": strings.TrimSpace(telemetry.ReadMode),
+		})
+	}
+	if telemetry.RAGFallbackUsed {
+		s.metrics.AddCounter(memoryobserve.MetricRAGFallbackTotal, 1, map[string]string{
+			"read_mode": strings.TrimSpace(telemetry.ReadMode),
+		})
+	}
+}
+
 // scoreRetrievedItem 计算 legacy 读链路的确定性排序分数。
 //
 // 说明：
--- a/backend/memory/service/retrieve_merge.go
+++ b/backend/memory/service/retrieve_merge.go
@@ -23,41 +23,50 @@ func (s *ReadService) HybridRetrieve(
 	effectiveSetting model.MemoryUserSetting,
 	limit int,
 	now time.Time,
-) ([]memorymodel.ItemDTO, error) {
+) ([]memorymodel.ItemDTO, retrieveTelemetry, error) {
+	telemetry := retrieveTelemetry{}
 	if s == nil || s.itemRepo == nil {
-		return nil, nil
+		return nil, telemetry, nil
 	}
 	if !effectiveSetting.MemoryEnabled {
-		return nil, nil
+		return nil, telemetry, nil
 	}

 	pinnedItems, err := s.retrievePinnedCandidates(ctx, req, effectiveSetting, now)
 	if err != nil {
-		return nil, err
+		return nil, telemetry, err
 	}
-	semanticItems, err := s.retrieveSemanticCandidates(ctx, req, effectiveSetting, limit, now)
+	telemetry.PinnedHitCount = len(pinnedItems)
+
+	semanticItems, semanticTelemetry, err := s.retrieveSemanticCandidates(ctx, req, effectiveSetting, limit, now)
 	if err != nil {
-		return nil, err
+		return nil, telemetry, err
 	}
+	telemetry.SemanticHitCount = len(semanticItems)
+	telemetry.Degraded = semanticTelemetry.Degraded
+	telemetry.RAGFallbackUsed = semanticTelemetry.RAGFallbackUsed

 	merged := make([]memorymodel.ItemDTO, 0, len(pinnedItems)+len(semanticItems))
 	merged = append(merged, pinnedItems...)
 	merged = append(merged, semanticItems...)
 	if len(merged) == 0 {
-		return nil, nil
+		return nil, telemetry, nil
 	}

+	beforeDedupCount := len(merged)
 	merged = dedupByID(merged)
 	merged = dedupByHash(merged)
 	merged = dedupByText(merged)
+	telemetry.DedupDropCount = beforeDedupCount - len(merged)
 	merged = RankItems(merged, now)
 	merged = applyTypeBudget(merged, s.cfg)
 	if len(merged) == 0 {
-		return nil, nil
+		return nil, telemetry, nil
 	}
+	telemetry.FinalCount = len(merged)

 	_ = s.itemRepo.TouchLastAccessAt(ctx, collectItemDTOIDs(merged), now)
-	return merged, nil
+	return merged, telemetry, nil
 }

 func (s *ReadService) retrievePinnedCandidates(
@@ -81,20 +90,26 @@ func (s *ReadService) retrieveSemanticCandidates(
 	effectiveSetting model.MemoryUserSetting,
 	limit int,
 	now time.Time,
-) ([]memorymodel.ItemDTO, error) {
+) ([]memorymodel.ItemDTO, semanticRetrieveTelemetry, error) {
+	telemetry := semanticRetrieveTelemetry{}
 	queryText := strings.TrimSpace(req.Query)
 	if queryText == "" {
-		return nil, nil
+		return nil, telemetry, nil
 	}

 	candidateLimit := hybridSemanticTopK(s.cfg, limit)
 	if s.cfg.RAGEnabled && s.ragRuntime != nil {
 		items, err := s.retrieveSemanticCandidatesByRAG(ctx, req, effectiveSetting, candidateLimit, now)
 		if shouldReturnSemanticRAGResult(items, err) {
-			return items, nil
+			telemetry.HitCount = len(items)
+			return items, telemetry, nil
 		}
+		telemetry.Degraded = true
+		telemetry.RAGFallbackUsed = true
 	}
-	return s.retrieveSemanticCandidatesByMySQL(ctx, req, effectiveSetting, candidateLimit, now)
+	items, err := s.retrieveSemanticCandidatesByMySQL(ctx, req, effectiveSetting, candidateLimit, now)
+	telemetry.HitCount = len(items)
+	return items, telemetry, err
 }

 func (s *ReadService) retrieveSemanticCandidatesByRAG(
--- a/backend/memory/utils/audit.go
+++ b/backend/memory/utils/audit.go
@@ -12,8 +12,12 @@ const (
 	AuditOperationCreate = "create"
 	// AuditOperationUpdate 表示决策层更新已有记忆的内容。
 	AuditOperationUpdate = "update"
+	// AuditOperationArchive 表示治理层把重复记忆归档。
+	AuditOperationArchive = "archive"
 	// AuditOperationDelete 表示对已有记忆做软删除。
 	AuditOperationDelete = "delete"
+	// AuditOperationRestore 表示把已删除/归档记忆恢复为 active。
+	AuditOperationRestore = "restore"
 )

 // BuildItemAuditLog 构造记忆变更审计日志。
--- a/backend/memory/vectorsync/syncer.go
+++ b/backend/memory/vectorsync/syncer.go
@@ -0,0 +1,213 @@
+package vectorsync
+
+import (
+	"context"
+	"fmt"
+	"log"
+	"strings"
+
+	infrarag "github.com/LoveLosita/smartflow/backend/infra/rag"
+	memoryobserve "github.com/LoveLosita/smartflow/backend/memory/observe"
+	memoryrepo "github.com/LoveLosita/smartflow/backend/memory/repo"
+	"github.com/LoveLosita/smartflow/backend/model"
+)
+
+// Syncer 负责 memory_items 与向量库之间的最小桥接。
+//
+// 职责边界：
+// 1. 只负责“把已经落库的记忆同步到 RAG / 从 RAG 删除”；
+// 2. 不负责决定哪些记忆该写、该删、该恢复，这些决策仍由上游 service/worker/cleanup 控制；
+// 3. 同步失败时只回写 vector_status 并打观测，不反向回滚业务事务，避免把在线链路拖成强依赖。
+type Syncer struct {
+	ragRuntime infrarag.Runtime
+	itemRepo   *memoryrepo.ItemRepo
+	observer   memoryobserve.Observer
+	metrics    memoryobserve.MetricsRecorder
+	logger     *log.Logger
+}
+
+func NewSyncer(
+	ragRuntime infrarag.Runtime,
+	itemRepo *memoryrepo.ItemRepo,
+	observer memoryobserve.Observer,
+	metrics memoryobserve.MetricsRecorder,
+) *Syncer {
+	if observer == nil {
+		observer = memoryobserve.NewNopObserver()
+	}
+	if metrics == nil {
+		metrics = memoryobserve.NewNopMetrics()
+	}
+	return &Syncer{
+		ragRuntime: ragRuntime,
+		itemRepo:   itemRepo,
+		observer:   observer,
+		metrics:    metrics,
+		logger:     log.Default(),
+	}
+}
+
+// Upsert 把新增/修改/恢复后的记忆同步到向量库。
+func (s *Syncer) Upsert(ctx context.Context, traceID string, items []model.MemoryItem) {
+	if s == nil || s.ragRuntime == nil || s.itemRepo == nil || len(items) == 0 {
+		return
+	}
+
+	requestItems := make([]infrarag.MemoryIngestItem, 0, len(items))
+	for _, item := range items {
+		requestItems = append(requestItems, infrarag.MemoryIngestItem{
+			MemoryID:         item.ID,
+			UserID:           item.UserID,
+			ConversationID:   strValue(item.ConversationID),
+			AssistantID:      strValue(item.AssistantID),
+			RunID:            strValue(item.RunID),
+			MemoryType:       item.MemoryType,
+			Title:            item.Title,
+			Content:          item.Content,
+			Confidence:       item.Confidence,
+			Importance:       item.Importance,
+			SensitivityLevel: item.SensitivityLevel,
+			IsExplicit:       item.IsExplicit,
+			Status:           item.Status,
+			TTLAt:            item.TTLAt,
+			CreatedAt:        item.CreatedAt,
+		})
+	}
+
+	result, err := s.ragRuntime.IngestMemory(memoryobserve.WithFields(ctx, map[string]any{
+		"trace_id": traceID,
+	}), infrarag.MemoryIngestRequest{
+		TraceID: traceID,
+		Action:  "add",
+		Items:   requestItems,
+	})
+	if err != nil {
+		s.observer.Observe(ctx, memoryobserve.Event{
+			Level:     memoryobserve.LevelWarn,
+			Component: memoryobserve.ComponentWrite,
+			Operation: "vector_upsert",
+			Fields: map[string]any{
+				"trace_id":   traceID,
+				"item_count": len(items),
+				"success":    false,
+				"error":      err,
+				"error_code": memoryobserve.ClassifyError(err),
+			},
+		})
+		for _, item := range items {
+			_ = s.itemRepo.UpdateVectorStateByID(ctx, item.ID, "failed", nil)
+		}
+		return
+	}
+
+	vectorIDMap := make(map[int64]string, len(result.DocumentIDs))
+	for _, documentID := range result.DocumentIDs {
+		memoryID := parseMemoryID(documentID)
+		if memoryID <= 0 {
+			continue
+		}
+		vectorIDMap[memoryID] = documentID
+	}
+
+	for _, item := range items {
+		vectorID := strPtrOrNil(vectorIDMap[item.ID])
+		_ = s.itemRepo.UpdateVectorStateByID(ctx, item.ID, "synced", vectorID)
+	}
+	s.observer.Observe(ctx, memoryobserve.Event{
+		Level:     memoryobserve.LevelInfo,
+		Component: memoryobserve.ComponentWrite,
+		Operation: "vector_upsert",
+		Fields: map[string]any{
+			"trace_id":       traceID,
+			"item_count":     len(items),
+			"document_count": len(result.DocumentIDs),
+			"success":        true,
+		},
+	})
+}
+
+// Delete 把一批记忆对应的向量从向量库中删除。
+func (s *Syncer) Delete(ctx context.Context, traceID string, memoryIDs []int64) {
+	if s == nil || len(memoryIDs) == 0 {
+		return
+	}
+	if s.ragRuntime == nil || s.itemRepo == nil {
+		return
+	}
+
+	documentIDs := make([]string, 0, len(memoryIDs))
+	for _, id := range memoryIDs {
+		documentIDs = append(documentIDs, fmt.Sprintf("memory:%d", id))
+	}
+
+	err := s.ragRuntime.DeleteMemory(memoryobserve.WithFields(ctx, map[string]any{
+		"trace_id": traceID,
+	}), documentIDs)
+	if err != nil {
+		s.observer.Observe(ctx, memoryobserve.Event{
+			Level:     memoryobserve.LevelWarn,
+			Component: memoryobserve.ComponentWrite,
+			Operation: "vector_delete",
+			Fields: map[string]any{
+				"trace_id":   traceID,
+				"item_count": len(memoryIDs),
+				"success":    false,
+				"error":      err,
+				"error_code": memoryobserve.ClassifyError(err),
+			},
+		})
+		for _, memoryID := range memoryIDs {
+			_ = s.itemRepo.UpdateVectorStateByID(ctx, memoryID, "failed", nil)
+		}
+		return
+	}
+
+	for _, memoryID := range memoryIDs {
+		_ = s.itemRepo.UpdateVectorStateByID(ctx, memoryID, "deleted", nil)
+	}
+	s.observer.Observe(ctx, memoryobserve.Event{
+		Level:     memoryobserve.LevelInfo,
+		Component: memoryobserve.ComponentWrite,
+		Operation: "vector_delete",
+		Fields: map[string]any{
+			"trace_id":   traceID,
+			"item_count": len(memoryIDs),
+			"success":    true,
+		},
+	})
+}
+
+func parseMemoryID(documentID string) int64 {
+	documentID = strings.TrimSpace(documentID)
+	if !strings.HasPrefix(documentID, "memory:") {
+		return 0
+	}
+	raw := strings.TrimPrefix(documentID, "memory:")
+	if strings.HasPrefix(raw, "uid:") {
+		return 0
+	}
+	var value int64
+	for _, ch := range raw {
+		if ch < '0' || ch > '9' {
+			return 0
+		}
+		value = value*10 + int64(ch-'0')
+	}
+	return value
+}
+
+func strPtrOrNil(v string) *string {
+	v = strings.TrimSpace(v)
+	if v == "" {
+		return nil
+	}
+	value := v
+	return &value
+}
+
+func strValue(v *string) string {
+	if v == nil {
+		return ""
+	}
+	return strings.TrimSpace(*v)
+}
--- a/backend/memory/worker/decision_flow.go
+++ b/backend/memory/worker/decision_flow.go
@@ -33,6 +33,11 @@ type factDecisionResult struct {
 	Outcomes []*ApplyActionOutcome
 }

+type candidateRecallResult struct {
+	Items        []memorymodel.CandidateSnapshot
+	FallbackMode string
+}
+
 // executeDecisionFlow 在 worker 内编排"召回→逐对比对→汇总→执行"全流程。
 //
 // 职责边界：
@@ -116,6 +121,7 @@ func (r *Runner) executeDecisionForFact(
 		}
 	}
 	if len(existing) > 0 {
+		r.recordDecisionObservation(ctx, job, payload, fact, 0, memorymodel.DecisionActionNone, "hash_exact", true, nil)
 		result.Outcomes = append(result.Outcomes, &ApplyActionOutcome{
 			Action:    memorymodel.DecisionActionNone,
 			NeedsSync: false,
@@ -124,7 +130,8 @@ func (r *Runner) executeDecisionForFact(
 	}

 	// Step 2: Milvus 语义召回（含降级）。
-	candidates := r.recallCandidates(ctx, payload, fact)
+	recallResult := r.recallCandidates(ctx, payload, fact)
+	candidates := recallResult.Items

 	// 打印召回候选详情，便于排查向量召回和阈值过滤效果。
 	if r.logger != nil {
@@ -151,9 +158,11 @@ func (r *Runner) executeDecisionForFact(
 	// Step 5: 校验 + 执行。
 	actionOutcome, err := ApplyFinalDecision(ctx, itemRepo, auditRepo, *decision, fact, job, payload)
 	if err != nil {
+		r.recordDecisionObservation(ctx, job, payload, fact, len(candidates), decision.Action, recallResult.FallbackMode, false, err)
 		return nil, fmt.Errorf("执行决策动作失败: %w", err)
 	}
 	result.Outcomes = append(result.Outcomes, actionOutcome)
+	r.recordDecisionObservation(ctx, job, payload, fact, len(candidates), decision.Action, recallResult.FallbackMode, true, nil)

 	// Step 6: conflict (DELETE) 后需要补一个 ADD 写入新 fact。
 	// 原因：旧记忆矛盾需删除，但新事实本身仍然有效，必须写入。
@@ -180,7 +189,7 @@ func (r *Runner) recallCandidates(
 	ctx context.Context,
 	payload memorymodel.ExtractJobPayload,
 	fact memorymodel.NormalizedFact,
-) []memorymodel.CandidateSnapshot {
+) candidateRecallResult {
 	// 1. 优先使用 Milvus 向量语义召回。
 	if r.ragRuntime != nil {
 		retrieveResult, err := r.ragRuntime.RetrieveMemory(ctx, infrarag.MemoryRetrieveRequest{
@@ -194,7 +203,10 @@ func (r *Runner) recallCandidates(
 		if err == nil && len(retrieveResult.Items) > 0 {
 			candidates := r.buildCandidatesFromRAG(retrieveResult.Items)
 			if len(candidates) > 0 {
-				return candidates
+				return candidateRecallResult{
+					Items:        candidates,
+					FallbackMode: "rag",
+				}
 			}
 			// RAG 返回了结果但 DocumentID 全部解析失败，降级到 MySQL。
 			if r.logger != nil {
@@ -204,10 +216,17 @@ func (r *Runner) recallCandidates(
 		if err != nil && r.logger != nil {
 			r.logger.Printf("[WARN][去重] Milvus 语义召回失败，降级到 MySQL: user_id=%d memory_type=%s topk=%d err=%v", payload.UserID, fact.MemoryType, r.cfg.DecisionCandidateTopK, err)
 		}
+		return candidateRecallResult{
+			Items:        r.recallCandidatesFromMySQL(ctx, payload, fact),
+			FallbackMode: "rag_to_mysql",
+		}
 	}

 	// 2. 降级：按 user_id + memory_type + status=active 查最近 N 条。
-	return r.recallCandidatesFromMySQL(ctx, payload, fact)
+	return candidateRecallResult{
+		Items:        r.recallCandidatesFromMySQL(ctx, payload, fact),
+		FallbackMode: "mysql_only",
+	}
 }

 // buildCandidatesFromRAG 从 RAG 检索结果构建候选快照列表。
--- a/backend/memory/worker/runner.go
+++ b/backend/memory/worker/runner.go
@@ -6,15 +6,16 @@ import (
 	"errors"
 	"fmt"
 	"log"
-	"strconv"
 	"strings"
 	"time"

 	infrarag "github.com/LoveLosita/smartflow/backend/infra/rag"
 	memorymodel "github.com/LoveLosita/smartflow/backend/memory/model"
+	memoryobserve "github.com/LoveLosita/smartflow/backend/memory/observe"
 	memoryorchestrator "github.com/LoveLosita/smartflow/backend/memory/orchestrator"
 	memoryrepo "github.com/LoveLosita/smartflow/backend/memory/repo"
 	memoryutils "github.com/LoveLosita/smartflow/backend/memory/utils"
+	memoryvectorsync "github.com/LoveLosita/smartflow/backend/memory/vectorsync"
 	"github.com/LoveLosita/smartflow/backend/model"
 	"gorm.io/gorm"
 )
@@ -42,6 +43,9 @@ type Runner struct {
 	extractor    Extractor
 	ragRuntime   infrarag.Runtime
 	logger       *log.Logger
+	vectorSyncer *memoryvectorsync.Syncer
+	observer     memoryobserve.Observer
+	metrics      memoryobserve.MetricsRecorder

 	// 决策层依赖。
 	// 说明：
@@ -62,7 +66,16 @@ func NewRunner(
 	ragRuntime infrarag.Runtime,
 	cfg memorymodel.Config,
 	decisionOrchestrator *memoryorchestrator.LLMDecisionOrchestrator,
+	vectorSyncer *memoryvectorsync.Syncer,
+	observer memoryobserve.Observer,
+	metrics memoryobserve.MetricsRecorder,
 ) *Runner {
+	if observer == nil {
+		observer = memoryobserve.NewNopObserver()
+	}
+	if metrics == nil {
+		metrics = memoryobserve.NewNopMetrics()
+	}
 	return &Runner{
 		db:                   db,
 		jobRepo:              jobRepo,
@@ -72,6 +85,9 @@ func NewRunner(
 		extractor:            extractor,
 		ragRuntime:           ragRuntime,
 		logger:               log.Default(),
+		vectorSyncer:         vectorSyncer,
+		observer:             observer,
+		metrics:              metrics,
 		cfg:                  cfg,
 		decisionOrchestrator: decisionOrchestrator,
 	}
@@ -96,6 +112,11 @@ func (r *Runner) RunOnce(ctx context.Context) (*RunOnceResult, error) {
 	if job == nil {
 		return &RunOnceResult{Claimed: false}, nil
 	}
+	if job.RetryCount > 0 {
+		r.metrics.AddCounter(memoryobserve.MetricJobRetryTotal, 1, map[string]string{
+			"job_type": strings.TrimSpace(job.JobType),
+		})
+	}

 	result := &RunOnceResult{
 		Claimed: true,
@@ -110,21 +131,25 @@ func (r *Runner) RunOnce(ctx context.Context) (*RunOnceResult, error) {
 		failReason := fmt.Sprintf("解析任务载荷失败: %v", err)
 		_ = r.jobRepo.MarkFailed(ctx, job.ID, failReason)
 		result.Status = model.MemoryJobStatusFailed
+		r.recordJobOutcome(ctx, job, nil, result.Status, false, err)
 		return result, nil
 	}

 	// 3. 先读取用户记忆设置。总开关关闭时，任务直接成功结束，不再继续抽取和落库。
 	setting, err := r.settingsRepo.GetByUserID(ctx, payload.UserID)
 	if err != nil {
+		r.recordJobOutcome(ctx, job, &payload, model.MemoryJobStatusFailed, false, err)
 		return nil, err
 	}
 	effectiveSetting := memoryutils.EffectiveUserSetting(setting, payload.UserID)
 	if !effectiveSetting.MemoryEnabled {
 		if err = r.jobRepo.MarkSuccess(ctx, job.ID); err != nil {
+			r.recordJobOutcome(ctx, job, &payload, model.MemoryJobStatusFailed, false, err)
 			return nil, err
 		}
 		result.Status = model.MemoryJobStatusSuccess
 		r.logger.Printf("memory worker skipped by user setting: job_id=%d user_id=%d", job.ID, payload.UserID)
+		r.recordJobOutcome(ctx, job, &payload, result.Status, true, nil)
 		return result, nil
 	}

@@ -134,26 +159,31 @@ func (r *Runner) RunOnce(ctx context.Context) (*RunOnceResult, error) {
 		failReason := fmt.Sprintf("抽取执行失败: %v", extractErr)
 		_ = r.jobRepo.MarkFailed(ctx, job.ID, failReason)
 		result.Status = model.MemoryJobStatusFailed
+		r.recordJobOutcome(ctx, job, &payload, result.Status, false, extractErr)
 		return result, nil
 	}
 	facts = memoryutils.FilterFactsBySetting(facts, effectiveSetting)

 	if len(facts) == 0 {
 		if err = r.jobRepo.MarkSuccess(ctx, job.ID); err != nil {
+			r.recordJobOutcome(ctx, job, &payload, model.MemoryJobStatusFailed, false, err)
 			return nil, err
 		}
 		result.Status = model.MemoryJobStatusSuccess
 		r.logger.Printf("memory worker run once noop: job_id=%d", job.ID)
+		r.recordJobOutcome(ctx, job, &payload, result.Status, true, nil)
 		return result, nil
 	}

 	items := buildMemoryItems(job, payload, facts)
 	if len(items) == 0 {
 		if err = r.jobRepo.MarkSuccess(ctx, job.ID); err != nil {
+			r.recordJobOutcome(ctx, job, &payload, model.MemoryJobStatusFailed, false, err)
 			return nil, err
 		}
 		result.Status = model.MemoryJobStatusSuccess
 		r.logger.Printf("memory worker run once empty-after-normalize: job_id=%d", job.ID)
+		r.recordJobOutcome(ctx, job, &payload, result.Status, true, nil)
 		return result, nil
 	}

@@ -169,16 +199,19 @@ func (r *Runner) RunOnce(ctx context.Context) (*RunOnceResult, error) {
 					failReason := fmt.Sprintf("决策降级后记忆落库失败: %v", err)
 					_ = r.jobRepo.MarkFailed(ctx, job.ID, failReason)
 					result.Status = model.MemoryJobStatusFailed
+					r.recordJobOutcome(ctx, job, &payload, result.Status, false, err)
 					return result, nil
 				}
 				result.Status = model.MemoryJobStatusSuccess
 				result.Facts = len(items)
 				r.syncMemoryVectors(ctx, items)
+				r.recordJobOutcome(ctx, job, &payload, result.Status, true, nil)
 				return result, nil
 			}
 			// FallbackMode=drop：丢弃本轮抽取结果，直接标记 job 成功。
 			_ = r.jobRepo.MarkSuccess(ctx, job.ID)
 			result.Status = model.MemoryJobStatusSuccess
+			r.recordJobOutcome(ctx, job, &payload, result.Status, true, nil)
 			return result, nil
 		}

@@ -189,6 +222,7 @@ func (r *Runner) RunOnce(ctx context.Context) (*RunOnceResult, error) {
 		r.syncVectorDeletes(ctx, outcome.VectorDeletes)
 		r.logger.Printf("[去重] 决策流程完成: job_id=%d user_id=%d 新增=%d 更新=%d 删除=%d 跳过=%d",
 			job.ID, payload.UserID, outcome.AddCount, outcome.UpdateCount, outcome.DeleteCount, outcome.NoneCount)
+		r.recordJobOutcome(ctx, job, &payload, result.Status, true, nil)
 		return result, nil
 	}

@@ -197,6 +231,7 @@ func (r *Runner) RunOnce(ctx context.Context) (*RunOnceResult, error) {
 		failReason := fmt.Sprintf("记忆落库失败: %v", err)
 		_ = r.jobRepo.MarkFailed(ctx, job.ID, failReason)
 		result.Status = model.MemoryJobStatusFailed
+		r.recordJobOutcome(ctx, job, &payload, result.Status, false, err)
 		return result, nil
 	}

@@ -204,6 +239,7 @@ func (r *Runner) RunOnce(ctx context.Context) (*RunOnceResult, error) {
 	result.Facts = len(items)
 	r.syncMemoryVectors(ctx, items)
 	r.logger.Printf("memory worker run once success: job_id=%d extracted_facts=%d", job.ID, len(items))
+	r.recordJobOutcome(ctx, job, &payload, result.Status, true, nil)
 	return result, nil
 }

@@ -268,56 +304,10 @@ func buildMemoryItems(job *model.MemoryJob, payload memorymodel.ExtractJobPayloa
 }

 func (r *Runner) syncMemoryVectors(ctx context.Context, items []model.MemoryItem) {
-	if r == nil || r.ragRuntime == nil || r.itemRepo == nil || len(items) == 0 {
+	if r == nil || r.vectorSyncer == nil || len(items) == 0 {
 		return
 	}
-
-	requestItems := make([]infrarag.MemoryIngestItem, 0, len(items))
-	for _, item := range items {
-		requestItems = append(requestItems, infrarag.MemoryIngestItem{
-			MemoryID:         item.ID,
-			UserID:           item.UserID,
-			ConversationID:   strValue(item.ConversationID),
-			AssistantID:      strValue(item.AssistantID),
-			RunID:            strValue(item.RunID),
-			MemoryType:       item.MemoryType,
-			Title:            item.Title,
-			Content:          item.Content,
-			Confidence:       item.Confidence,
-			Importance:       item.Importance,
-			SensitivityLevel: item.SensitivityLevel,
-			IsExplicit:       item.IsExplicit,
-			Status:           item.Status,
-			TTLAt:            item.TTLAt,
-			CreatedAt:        item.CreatedAt,
-		})
-	}
-
-	result, err := r.ragRuntime.IngestMemory(ctx, infrarag.MemoryIngestRequest{
-		Action: "add",
-		Items:  requestItems,
-	})
-	if err != nil {
-		r.logger.Printf("[WARN][去重] 记忆向量同步失败: count=%d err=%v", len(items), err)
-		for _, item := range items {
-			_ = r.itemRepo.UpdateVectorStateByID(ctx, item.ID, "failed", nil)
-		}
-		return
-	}
-
-	vectorIDMap := make(map[int64]string, len(result.DocumentIDs))
-	for _, documentID := range result.DocumentIDs {
-		memoryID := parseMemoryID(documentID)
-		if memoryID <= 0 {
-			continue
-		}
-		vectorIDMap[memoryID] = documentID
-	}
-
-	for _, item := range items {
-		vectorID := strPtrOrNil(vectorIDMap[item.ID])
-		_ = r.itemRepo.UpdateVectorStateByID(ctx, item.ID, "synced", vectorID)
-	}
+	r.vectorSyncer.Upsert(ctx, "", items)
 }

 // syncVectorDeletes 处理决策层 DELETE 动作产出的向量清理需求。
@@ -327,33 +317,10 @@ func (r *Runner) syncMemoryVectors(ctx context.Context, items []model.MemoryItem
 // 2. 调 Runtime.DeleteMemory 真正从 Milvus 删除对应向量；
 // 3. 更新 MySQL vector_status 标记删除结果。
 func (r *Runner) syncVectorDeletes(ctx context.Context, memoryIDs []int64) {
-	if r == nil || len(memoryIDs) == 0 {
+	if r == nil || r.vectorSyncer == nil || len(memoryIDs) == 0 {
 		return
 	}
-
-	// 1. 构造 documentID 列表。
-	documentIDs := make([]string, 0, len(memoryIDs))
-	for _, id := range memoryIDs {
-		documentIDs = append(documentIDs, fmt.Sprintf("memory:%d", id))
-	}
-
-	// 2. 调 Runtime 删除向量。
-	if r.ragRuntime != nil {
-		if err := r.ragRuntime.DeleteMemory(ctx, documentIDs); err != nil {
-			r.logger.Printf("[WARN][去重] Milvus 向量删除失败，标记为 pending 等待后续清理: count=%d ids=%v err=%v", len(memoryIDs), memoryIDs, err)
-		} else {
-			r.logger.Printf("[去重] Milvus 向量删除完成: count=%d ids=%v", len(memoryIDs), memoryIDs)
-		}
-	}
-
-	// 3. 更新 MySQL vector_status。
-	for _, memoryID := range memoryIDs {
-		if updateErr := r.itemRepo.UpdateVectorStateByID(ctx, memoryID, "deleted", nil); updateErr != nil {
-			if r.logger != nil {
-				r.logger.Printf("[WARN] 向量状态更新失败: memory_id=%d err=%v", memoryID, updateErr)
-			}
-		}
-	}
+	r.vectorSyncer.Delete(ctx, "", memoryIDs)
 }

 func resolveMemoryTTLAt(base time.Time, memoryType string) *time.Time {
@@ -395,11 +362,106 @@ func int64PtrOrNil(v int64) *int64 {
 	return &value
 }

-func strValue(v *string) string {
-	if v == nil {
-		return ""
+func (r *Runner) recordJobOutcome(
+	ctx context.Context,
+	job *model.MemoryJob,
+	payload *memorymodel.ExtractJobPayload,
+	status string,
+	success bool,
+	err error,
+) {
+	if r == nil {
+		return
 	}
-	return strings.TrimSpace(*v)
+
+	level := memoryobserve.LevelInfo
+	if !success || err != nil {
+		level = memoryobserve.LevelWarn
+	}
+	fields := map[string]any{
+		"job_id":     jobIDValue(job),
+		"status":     strings.TrimSpace(status),
+		"success":    success && err == nil,
+		"error":      err,
+		"error_code": memoryobserve.ClassifyError(err),
+	}
+	if payload != nil {
+		fields["trace_id"] = strings.TrimSpace(payload.TraceID)
+		fields["user_id"] = payload.UserID
+		fields["conversation_id"] = strings.TrimSpace(payload.ConversationID)
+	}
+
+	r.observer.Observe(ctx, memoryobserve.Event{
+		Level:     level,
+		Component: memoryobserve.ComponentWrite,
+		Operation: "job",
+		Fields:    fields,
+	})
+	r.metrics.AddCounter(memoryobserve.MetricJobTotal, 1, map[string]string{
+		"status": strings.TrimSpace(status),
+	})
+}
+
+func (r *Runner) recordDecisionObservation(
+	ctx context.Context,
+	job *model.MemoryJob,
+	payload memorymodel.ExtractJobPayload,
+	fact memorymodel.NormalizedFact,
+	candidateCount int,
+	finalAction string,
+	fallbackMode string,
+	success bool,
+	err error,
+) {
+	if r == nil {
+		return
+	}
+
+	level := memoryobserve.LevelInfo
+	status := "success"
+	if !success || err != nil {
+		level = memoryobserve.LevelWarn
+		status = "error"
+	}
+	fallbackMode = strings.TrimSpace(fallbackMode)
+	if fallbackMode == "" {
+		fallbackMode = "none"
+	}
+
+	r.observer.Observe(ctx, memoryobserve.Event{
+		Level:     level,
+		Component: memoryobserve.ComponentWrite,
+		Operation: memoryobserve.OperationDecision,
+		Fields: map[string]any{
+			"trace_id":        strings.TrimSpace(payload.TraceID),
+			"user_id":         payload.UserID,
+			"conversation_id": strings.TrimSpace(payload.ConversationID),
+			"job_id":          jobIDValue(job),
+			"fact_type":       strings.TrimSpace(fact.MemoryType),
+			"candidate_count": candidateCount,
+			"final_action":    strings.TrimSpace(finalAction),
+			"fallback_mode":   fallbackMode,
+			"success":         success && err == nil,
+			"error":           err,
+			"error_code":      memoryobserve.ClassifyError(err),
+		},
+	})
+	r.metrics.AddCounter(memoryobserve.MetricDecisionTotal, 1, map[string]string{
+		"action": strings.TrimSpace(finalAction),
+		"status": status,
+	})
+	if fallbackMode != "none" && fallbackMode != "hash_exact" && fallbackMode != "rag" {
+		r.metrics.AddCounter(memoryobserve.MetricDecisionFallbackTotal, 1, map[string]string{
+			"mode": fallbackMode,
+		})
+	}
+}
+
+func jobIDValue(job *model.MemoryJob) int64 {
+	if job == nil {
+		return 0
+	}
+	return job.ID
 }

 func parseMemoryID(documentID string) int64 {
@@ -411,9 +473,13 @@ func parseMemoryID(documentID string) int64 {
 	if strings.HasPrefix(raw, "uid:") {
 		return 0
 	}
-	memoryID, err := strconv.ParseInt(raw, 10, 64)
-	if err != nil {
+
+	var value int64
+	for _, ch := range raw {
+		if ch < '0' || ch > '9' {
 			return 0
 		}
-	return memoryID
+		value = value*10 + int64(ch-'0')
+	}
+	return value
 }
--- a/backend/model/memory_manage.go
+++ b/backend/model/memory_manage.go
@@ -0,0 +1,105 @@
+package model
+
+import "time"
+
+// MemoryGetItemRequest 描述“查看我的某条记忆”所需的最小参数。
+type MemoryGetItemRequest struct {
+	UserID   int
+	MemoryID int64
+}
+
+// MemoryCreateItemRequest 描述“手动新增一条记忆”的输入。
+type MemoryCreateItemRequest struct {
+	UserID           int        `json:"-"`
+	ConversationID   string     `json:"conversation_id,omitempty"`
+	AssistantID      string     `json:"assistant_id,omitempty"`
+	RunID            string     `json:"run_id,omitempty"`
+	MemoryType       string     `json:"memory_type"`
+	Title            string     `json:"title"`
+	Content          string     `json:"content"`
+	Confidence       *float64   `json:"confidence,omitempty"`
+	Importance       *float64   `json:"importance,omitempty"`
+	SensitivityLevel *int       `json:"sensitivity_level,omitempty"`
+	IsExplicit       *bool      `json:"is_explicit,omitempty"`
+	TTLAt            *time.Time `json:"ttl_at,omitempty"`
+	Reason           string     `json:"reason,omitempty"`
+	OperatorType     string     `json:"-"`
+}
+
+// MemoryUpdateItemRequest 描述“手动修改一条记忆”的 Patch 输入。
+//
+// 说明：
+// 1. 使用指针区分“未传字段”和“显式传零值”；
+// 2. ClearTTL 用于表达“显式清空 ttl_at”；
+// 3. 当前仍只允许修改内容侧字段，不开放跨用户、跨归属字段改写。
+type MemoryUpdateItemRequest struct {
+	UserID           int        `json:"-"`
+	MemoryID         int64      `json:"-"`
+	MemoryType       *string    `json:"memory_type,omitempty"`
+	Title            *string    `json:"title,omitempty"`
+	Content          *string    `json:"content,omitempty"`
+	Confidence       *float64   `json:"confidence,omitempty"`
+	Importance       *float64   `json:"importance,omitempty"`
+	SensitivityLevel *int       `json:"sensitivity_level,omitempty"`
+	IsExplicit       *bool      `json:"is_explicit,omitempty"`
+	TTLAt            *time.Time `json:"ttl_at,omitempty"`
+	ClearTTL         bool       `json:"clear_ttl,omitempty"`
+	Reason           string     `json:"reason,omitempty"`
+	OperatorType     string     `json:"-"`
+}
+
+// MemoryDeleteItemRequest 描述“删除我的一条记忆”的输入。
+type MemoryDeleteItemRequest struct {
+	UserID       int
+	MemoryID     int64
+	Reason       string
+	OperatorType string
+}
+
+// MemoryRestoreItemRequest 描述“恢复我的一条记忆”的输入。
+type MemoryRestoreItemRequest struct {
+	UserID       int
+	MemoryID     int64
+	Reason       string
+	OperatorType string
+}
+
+// MemoryDedupCleanupRequest 描述离线去重治理任务的执行参数。
+type MemoryDedupCleanupRequest struct {
+	UserID       int
+	Limit        int
+	DryRun       bool
+	Reason       string
+	OperatorType string
+}
+
+// MemoryDedupCleanupResult 描述一次离线去重治理的汇总结果。
+type MemoryDedupCleanupResult struct {
+	ScannedGroupCount int     `json:"scanned_group_count"`
+	DedupedGroupCount int     `json:"deduped_group_count"`
+	KeptCount         int     `json:"kept_count"`
+	ArchivedCount     int     `json:"archived_count"`
+	ArchivedIDs       []int64 `json:"archived_ids,omitempty"`
+	DryRun            bool    `json:"dry_run"`
+}
+
+// MemoryItemView 是前端可见的记忆条目视图。
+type MemoryItemView struct {
+	ID               int64      `json:"id"`
+	UserID           int        `json:"user_id"`
+	ConversationID   string     `json:"conversation_id,omitempty"`
+	AssistantID      string     `json:"assistant_id,omitempty"`
+	RunID            string     `json:"run_id,omitempty"`
+	MemoryType       string     `json:"memory_type"`
+	Title            string     `json:"title"`
+	Content          string     `json:"content"`
+	ContentHash      string     `json:"content_hash,omitempty"`
+	Confidence       float64    `json:"confidence"`
+	Importance       float64    `json:"importance"`
+	SensitivityLevel int        `json:"sensitivity_level"`
+	IsExplicit       bool       `json:"is_explicit"`
+	Status           string     `json:"status"`
+	TTLAt            *time.Time `json:"ttl_at,omitempty"`
+	CreatedAt        *time.Time `json:"created_at,omitempty"`
+	UpdatedAt        *time.Time `json:"updated_at,omitempty"`
+}
--- a/backend/respond/respond.go
+++ b/backend/respond/respond.go
@@ -339,6 +339,21 @@ var ( //请求相关的响应
 		Info:   "conversation_id is required when confirm_action is present",
 	}

+	MemoryItemNotFound = Response{ //记忆条目不存在
+		Status: "40055",
+		Info:   "memory item not found",
+	}
+
+	MemoryInvalidType = Response{ //记忆类型不合法
+		Status: "40056",
+		Info:   "invalid memory type",
+	}
+
+	MemoryInvalidContent = Response{ //记忆内容为空或不合法
+		Status: "40057",
+		Info:   "invalid memory content",
+	}
+
 	RouteControlInternalError = Response{ //路由控制码内部错误
 		Status: "50001",
 		Info:   "route control failed",
--- a/backend/routers/routers.go
+++ b/backend/routers/routers.go
@@ -97,6 +97,16 @@ func RegisterRouters(handlers *api.ApiHandlers, cache *dao.CacheDAO, userRepo *d
 			agentGroup.GET("/schedule-preview", handlers.AgentHandler.GetSchedulePlanPreview)
 			agentGroup.GET("/context-stats", handlers.AgentHandler.GetContextStats)
 		}
+		memoryGroup := apiGroup.Group("/memory")
+		{
+			memoryGroup.Use(middleware.JWTTokenAuth(cache), middleware.RateLimitMiddleware(limiter, 20, 1))
+			memoryGroup.GET("/items", handlers.MemoryHandler.ListItems)
+			memoryGroup.GET("/items/:id", handlers.MemoryHandler.GetItem)
+			memoryGroup.POST("/items", middleware.IdempotencyMiddleware(cache), handlers.MemoryHandler.CreateItem)
+			memoryGroup.PATCH("/items/:id", middleware.IdempotencyMiddleware(cache), handlers.MemoryHandler.UpdateItem)
+			memoryGroup.DELETE("/items/:id", middleware.IdempotencyMiddleware(cache), handlers.MemoryHandler.DeleteItem)
+			memoryGroup.POST("/items/:id/restore", middleware.IdempotencyMiddleware(cache), handlers.MemoryHandler.RestoreItem)
+		}
 	}
 	// 初始化Gin引擎
 	log.Println("Routes setup completed")
--- a/backend/service/agentsvc/agent.go
+++ b/backend/service/agentsvc/agent.go
@@ -16,6 +16,7 @@ import (
 	outboxinfra "github.com/LoveLosita/smartflow/backend/infra/outbox"
 	"github.com/LoveLosita/smartflow/backend/inits"
 	memorymodel "github.com/LoveLosita/smartflow/backend/memory/model"
+	memoryobserve "github.com/LoveLosita/smartflow/backend/memory/observe"
 	"github.com/LoveLosita/smartflow/backend/model"
 	newagentmodel "github.com/LoveLosita/smartflow/backend/newAgent/model"
 	newagenttools "github.com/LoveLosita/smartflow/backend/newAgent/tools"
@@ -60,6 +61,8 @@ type AgentService struct {
 	compactionStore   newagentmodel.CompactionStore
 	memoryReader      MemoryReader
 	memoryCfg         memorymodel.Config
+	memoryObserver    memoryobserve.Observer
+	memoryMetrics     memoryobserve.MetricsRecorder
 }

 // NewAgentService 构造 AgentService。
--- a/backend/service/agentsvc/agent_memory.go
+++ b/backend/service/agentsvc/agent_memory.go
@@ -7,6 +7,7 @@ import (
 	"time"

 	memorymodel "github.com/LoveLosita/smartflow/backend/memory/model"
+	memoryobserve "github.com/LoveLosita/smartflow/backend/memory/observe"
 	newagentmodel "github.com/LoveLosita/smartflow/backend/newAgent/model"
 )

@@ -27,10 +28,21 @@ type MemoryReader interface {
 	Retrieve(ctx context.Context, req memorymodel.RetrieveRequest) ([]memorymodel.ItemDTO, error)
 }

+type memoryObserveProvider interface {
+	MemoryObserver() memoryobserve.Observer
+	MemoryMetrics() memoryobserve.MetricsRecorder
+}
+
 // SetMemoryReader 注入 newAgent 主链路读取记忆所需的薄接口与渲染配置。
 func (s *AgentService) SetMemoryReader(reader MemoryReader, cfg memorymodel.Config) {
 	s.memoryReader = reader
 	s.memoryCfg = cfg
+	s.memoryObserver = memoryobserve.NewNopObserver()
+	s.memoryMetrics = memoryobserve.NewNopMetrics()
+	if provider, ok := reader.(memoryObserveProvider); ok {
+		s.memoryObserver = provider.MemoryObserver()
+		s.memoryMetrics = provider.MemoryMetrics()
+	}
 }

 // injectMemoryContext 在 graph 执行前，把本轮相关记忆写入 ConversationContext 的 pinned block。
@@ -64,6 +76,7 @@ func (s *AgentService) injectMemoryContext(
 	})
 	if err != nil {
 		conversationContext.RemovePinnedBlock(newAgentMemoryBlockKey)
+		s.recordMemoryInject(ctx, userID, 0, false, err)
 		log.Printf("读取记忆上下文失败 user=%d chat=%s err=%v", userID, chatID, err)
 		return
 	}
@@ -71,6 +84,7 @@ func (s *AgentService) injectMemoryContext(
 	content := renderMemoryPinnedContentByMode(items, s.memoryCfg.EffectiveInjectRenderMode())
 	if content == "" {
 		conversationContext.RemovePinnedBlock(newAgentMemoryBlockKey)
+		s.recordMemoryInject(ctx, userID, len(items), false, nil)
 		return
 	}

@@ -79,6 +93,7 @@ func (s *AgentService) injectMemoryContext(
 		Title:   newAgentMemoryBlockTitle,
 		Content: content,
 	})
+	s.recordMemoryInject(ctx, userID, len(items), true, nil)
 }

 // shouldInjectMemoryForInput 判断当前输入是否值得触发一次记忆召回。
@@ -100,3 +115,49 @@ func shouldInjectMemoryForInput(userMessage string) bool {
 		return true
 	}
 }
+
+func (s *AgentService) recordMemoryInject(
+	ctx context.Context,
+	userID int,
+	inputCount int,
+	success bool,
+	err error,
+) {
+	if s == nil {
+		return
+	}
+	observer := s.memoryObserver
+	if observer == nil {
+		observer = memoryobserve.NewNopObserver()
+	}
+	metrics := s.memoryMetrics
+	if metrics == nil {
+		metrics = memoryobserve.NewNopMetrics()
+	}
+
+	level := memoryobserve.LevelInfo
+	if err != nil {
+		level = memoryobserve.LevelWarn
+	}
+	observer.Observe(ctx, memoryobserve.Event{
+		Level:     level,
+		Component: memoryobserve.ComponentInject,
+		Operation: memoryobserve.OperationInject,
+		Fields: map[string]any{
+			"user_id":        userID,
+			"inject_mode":    s.memoryCfg.EffectiveInjectRenderMode(),
+			"input_count":    inputCount,
+			"rendered_count": inputCount,
+			"token_budget":   0,
+			"fallback":       false,
+			"success":        success && err == nil,
+			"error":          err,
+			"error_code":     memoryobserve.ClassifyError(err),
+		},
+	})
+	if inputCount > 0 {
+		metrics.AddCounter(memoryobserve.MetricInjectItemTotal, int64(inputCount), map[string]string{
+			"inject_mode": s.memoryCfg.EffectiveInjectRenderMode(),
+		})
+	}
+}