跳到主要内容
版本:v0.3

智能体记忆(Agentic Memory)

执行摘要​

本文描述 Semantic Router 中智能体记忆的概念验证(POC)。智能体记忆使 AI 智能体能够跨会话记住信息,从而提供连续性与个性化。

POC 范围: 本文为概念验证,非生产级设计。目标是验证核心记忆流(检索 → 注入 → 提取 → 存储)在可接受准确度下可行。生产加固(错误处理、扩展、监控)不在范围内。

核心能力​

能力说明
记忆检索基于嵌入的检索与简单预过滤
记忆写入基于 LLM 的事实与流程提取
跨会话持久化记忆存于 Milvus(重启可保留;生产级备份/高可用未验证)
用户隔离按 user_id 划分(见下表)

用户隔离与 Milvus 性能说明:

方式POC生产(1 万+ 用户)
简单过滤检索后按 user_id 过滤退化:先搜全库再过滤
分区键POC 过重物理隔离,每用户 O(log N)
标量索引POC 过重对 user_id 建索引以加速过滤

POC: 使用简单元数据过滤(测试足够)。
生产: 在 Milvus schema 中将 user_id 配为分区键或标量索引字段。

关键设计原则​

  1. 简单预过滤 决定是否检索记忆
  2. 利用历史 上下文窗口 对查询消歧
  3. LLM 提取事实 并在保存时分类
  4. 对检索结果做 基于阈值的过滤

POC 明确假设​

假设含义若错误的风险
LLM 提取基本准确可能存入错误事实记忆污染(可用 Forget API 修复)
0.6 相似度阈值为起点可能需调参可依检索质量日志调整
Milvus 可用且已配置宕机则功能关闭优雅降级(不崩溃)
嵌入模型输出 384 维向量须与 Milvus schema 一致启动失败(可检测)
可通过 Response API 链获得历史上下文所需无历史则跳过记忆

目录​

  1. 问题陈述
  2. 架构概览
  3. 记忆类型
  4. 流水线集成
  5. 记忆检索
  6. 记忆写入
  7. 记忆操作
  8. 数据结构
  9. API 扩展
  10. 配置
  11. 失败模式与回退(POC)
  12. 成功标准(POC)
  13. 实现计划
  14. 后续增强

1. 问题陈述​

现状​

Response API 通过 previous_response_id 提供会话链,但跨会话知识会丢失:

Session A (March 15):
User: "My budget for the Hawaii trip is $10,000"
→ Saved in session chain

Session B (March 20) - NEW SESSION:
User: "What's my budget for the trip?"
→ No previous_response_id → Knowledge LOST ❌

目标状态​

使用智能体记忆时:

Session A (March 15):
User: "My budget for the Hawaii trip is $10,000"
→ Extracted and saved to Milvus

Session B (March 20) - NEW SESSION:
User: "What's my budget for the trip?"
→ Pre-filter: memory-relevant ✓
→ Search Milvus → Found: "budget for Hawaii is $10K"
→ Inject into LLM context
→ Assistant: "Your budget for the Hawaii trip is $10,000!" ✅

2. 架构概览​

┌─────────────────────────────────────────────────────────────────────────┐
│ AGENTIC MEMORY ARCHITECTURE │
├─────────────────────────────────────────────────────────────────────────┤
│ │
│ ExtProc Pipeline │
│ ┌──────────────────────────────────────────────────────────────────┐ │
│ │ │ │
│ │ Request → Fact? → Tool? → Security → Cache → MEMORY → LLM │ │
│ │ │ │ ↑↓ │ │
│ │ └───────┴──── signals used ────────┘ │ │
│ │ │ │
│ │ Response ← [extract & store] ←─────────────────┘ │ │
│ │ │ │
│ └──────────────────────────────────────────────────────────────────┘ │
│ │ │
│ ┌─────────────────────┴─────────────────────┐ │
│ │ │ │
│ ┌─────────▼─────────┐ ┌────────────▼───┐ │
│ │ Memory Retrieval │ │ Memory Saving │ │
│ │ (request phase) │ │(response phase)│ │
│ ├───────────────────┤ ├────────────────┤ │
│ │ 1. Check signals │ │ 1. LLM extract │ │
│ │ (Fact? Tool?) │ │ 2. Classify │ │
│ │ 2. Build context │ │ 3. Deduplicate │ │
│ │ 3. Milvus search │ │ 4. Store │ │
│ │ 4. Inject to LLM │ │ │ │
│ └─────────┬─────────┘ └────────┬───────┘ │
│ │ │ │
│ │ ┌──────────────┐ │ │
│ └────────►│ Milvus │◄─────────────┘ │
│ └──────────────┘ │
│ │
└─────────────────────────────────────────────────────────────────────────┘

组件职责​

组件职责位置
Memory Filter决策 + 检索 + 注入pkg/extproc/req_filter_memory.go
Memory Extractor基于 LLM 的事实提取pkg/memory/extractor.go(新建)
Memory Store存储接口pkg/memory/store.go
Milvus Store向量库后端pkg/memory/milvus_store.go
Existing ClassifiersFact/Tool 信号(复用)pkg/extproc/processor_req_body.go

存储架构​

Issue #808 提出了多层存储架构。本文分阶段实现:

┌─────────────────────────────────────────────────────────────────────────┐
│ STORAGE ARCHITECTURE (Phased) │
├─────────────────────────────────────────────────────────────────────────┤
│ │
│ ┌─────────────────────────────────────────────────────────────────┐ │
│ │ PHASE 1 (MVP) │ │
│ │ ┌─────────────────────────────────────────────────────────┐ │ │
│ │ │ Milvus (Vector Index) │ │ │
│ │ │ • Semantic search over memories │ │ │
│ │ │ • Embedding storage │ │ │
│ │ │ • Content + metadata │ │ │
│ │ └─────────────────────────────────────────────────────────┘ │ │
│ └─────────────────────────────────────────────────────────────────┘ │
│ │
│ ┌─────────────────────────────────────────────────────────────────┐ │
│ │ PHASE 2 (Performance) │ │
│ │ ┌─────────────────────────────────────────────────────────┐ │ │
│ │ │ Redis (Hot Cache) │ │ │
│ │ │ • Fast metadata lookup │ │ │
│ │ │ • Recently accessed memories │ │ │
│ │ │ • TTL/expiration support │ │ │
│ │ └─────────────────────────────────────────────────────────┘ │ │
│ └─────────────────────────────────────────────────────────────────┘ │
│ │
│ ┌─────────────────────────────────────────────────────────────────┐ │
│ │ PHASE 3+ (If Needed) │ │
│ │ ┌───────────────────────┐ ┌───────────────────────┐ │ │
│ │ │ Graph Store (Neo4j) │ │ Time-Series Index │ │ │
│ │ │ • Memory links │ │ • Temporal queries │ │ │
│ │ │ • Relationships │ │ • Decay scoring │ │ │
│ │ └───────────────────────┘ └───────────────────────┘ │ │
│ └─────────────────────────────────────────────────────────────────┘ │
│ │
└─────────────────────────────────────────────────────────────────────────┘
层用途何时需要状态
Milvus语义向量检索核心能力MVP
Redis热缓存、快速访问、TTL性能优化Phase 2
Graph (Neo4j)记忆关联多跳推理查询按需
Time-Series时序查询、衰减按时间的重要性打分按需

设计决策: 先从仅 Milvus 开始。其余层按实证需求增加,而非臆测。Store 接口抽象存储,后续可换后端而不改检索/写入逻辑。


3. 记忆类型​

类型用途示例状态
Semantic事实、偏好、知识"User's budget for Hawaii is $10,000"MVP
Procedural步骤、流程"To deploy payment-service: run npm build, then docker push"MVP
Episodic会话摘要、过往事件"On Dec 29 2024, user planned Hawaii vacation with $10K budget"MVP(受限)
Reflective自省、经验教训"Previous budget response was incomplete - user prefers detailed breakdowns"未来

情景记忆(MVP 限制): 未实现会话结束检测。情景记忆仅在 LLM 提取显式产出摘要式内容时创建。可靠的会话结束触发推迟到 Phase 2。

反思记忆: 自省与经验教训。不在本 POC 范围内。见 附录 A。

记忆向量空间​

记忆按内容/主题聚类,而非按类型。类型是元数据:

┌────────────────────────────────────────────────────────────────────────┐
│ MEMORY VECTOR SPACE │
│ │
│ ┌─────────────────┐ ┌─────────────────┐ │
│ │ BUDGET/MONEY │ │ DEPLOYMENT │ │
│ │ CLUSTER │ │ CLUSTER │ │
│ │ │ │ │ │
│ │ ● budget=$10K │ │ ● npm build │ │
│ │ (semantic) │ │ (procedural) │ │
│ │ ● cost=$5K │ │ ● docker push │ │
│ │ (semantic) │ │ (procedural) │ │
│ └─────────────────┘ └─────────────────┘ │
│ │
│ ● = memory with type as metadata │
│ Query matches content → type comes from matched memory │
│ │
└────────────────────────────────────────────────────────────────────────┘

Response API 与智能体记忆:何时有价值?​

关键区分: 当存在 previous_response_id 时,Response API 已把完整对话历史发给 LLM。智能体记忆的价值在于跨会话上下文。

┌─────────────────────────────────────────────────────────────────────────┐
│ RESPONSE API vs. AGENTIC MEMORY: CONTEXT SOURCES │
├─────────────────────────────────────────────────────────────────────────┤
│ │
│ SAME SESSION (has previous_response_id): │
│ ───────────────────────────────────────── │
│ Response API provides: │
│ └── Full conversation chain (all turns) → sent to LLM │
│ │
│ Agentic Memory: │
│ └── STILL VALUABLE - current session may not have the answer │
│ └── Example: 100 turns planning vacation, but budget never said │
│ └── Days ago: "I have 10K spare, is that enough for a week in │
│ Thailand?" → LLM extracts: "User has $10K budget for trip" │
│ └── Now: "What's my budget?" → answer in memory, not this chain │
│ │
│ NEW SESSION (no previous_response_id): │
│ ────────────────────────────────────── │
│ Response API provides: │
│ └── Nothing (no chain to follow) │
│ │
│ Agentic Memory: │
│ └── ADDS VALUE - retrieves cross-session context │
│ └── "What was my Hawaii budget?" → finds fact from March session │
│ │
└─────────────────────────────────────────────────────────────────────────┘

设计决策: 记忆检索在两种场景下都有价值——新会话(无链)与已有会话(查询可能引用其他会话)。预过滤通过时始终检索。

已知冗余: 若答案已在当前链中,仍会检索记忆(浪费约 10–30ms)。若不语义理解查询,无法廉价判断「答案是否已在历史中」。POC 接受该开销。

Phase 2 方案: 上下文压缩 可正确处理——不再由 Response API 发送全量历史,而发送压缩摘要 + 最近轮次 + 相关记忆。摘要在汇总时提取事实,从而消除冗余。


4. 流水线集成​

当前流水线(main 分支)​

1. Response API Translation
2. Parse Request
3. Fact-Check Classification
4. Tool Detection
5. Decision & Model Selection
6. Security Checks
7. PII Detection
8. Semantic Cache Check
9. Model Routing → LLM

集成智能体记忆后的增强流水线​

REQUEST PHASE:
─────────────
1. Response API Translation
2. Parse Request
3. Fact-Check Classification ──┐
4. Tool Detection ├── Existing signals
5. Decision & Model Selection ──┘
6. Security Checks
7. PII Detection
8. Semantic Cache Check ───► if HIT → return cached
9. 🆕 Memory Decision:
└── if (NOT Fact) AND (NOT Tool) AND (NOT Greeting) → continue
└── else → skip to step 12
10. 🆕 Build context + rewrite query [~1-5ms]
11. 🆕 Search Milvus, inject memories [~10-30ms]
12. Model Routing → LLM

RESPONSE PHASE:
──────────────
13. Parse LLM Response
14. Cache Update
15. 🆕 Memory Extraction (async goroutine, if auto_store enabled)
└── Runs in background, does NOT add latency to response
16. Response API Translation
17. Return to Client

第 10 步说明: 查询改写策略(上下文前缀、LLM 改写、HyDE)见 附录 C。


5. 记忆检索​

流程​

┌─────────────────────────────────────────────────────────────────────────┐
│ MEMORY RETRIEVAL FLOW │
├─────────────────────────────────────────────────────────────────────────┤
│ │
│ 1. MEMORY DECISION (reuse existing pipeline signals) │
│ ────────────────────────────────────────────────── │
│ │
│ Pipeline already classified: │
│ ├── ctx.IsFact (Fact-Check classifier) │
│ ├── ctx.RequiresTool (Tool Detection) │
│ └── isGreeting(query) (simple pattern) │
│ │
│ Decision: │
│ ├── Fact query? → SKIP (general knowledge) │
│ ├── Tool query? → SKIP (tool provides answer) │
│ ├── Greeting? → SKIP (no context needed) │
│ └── Otherwise → SEARCH MEMORY │
│ │
│ 2. BUILD CONTEXT + REWRITE QUERY │
│ ───────────────────────────── │
│ History: ["Planning vacation", "Hawaii sounds nice"] │
│ Query: "How much?" │
│ │
│ Option A (MVP): Context prepend │
│ → "How much? Hawaii vacation planning" │
│ │
│ Option B (v1): LLM rewrite │
│ → "What is the budget for the Hawaii vacation?" │
│ │
│ 3. MILVUS SEARCH │
│ ───────────── │
│ Embed context → Search with user_id filter → Top-k results │
│ │
│ 4. THRESHOLD FILTER │
│ ──────────────── │
│ Keep only results with similarity > 0.6 │
│ ⚠️ Threshold is configurable; 0.6 is starting value, tune via logs │
│ │
│ 5. INJECT INTO LLM CONTEXT │
│ ──────────────────────── │
│ Add as system message: "User's relevant context: ..." │
│ │
└─────────────────────────────────────────────────────────────────────────┘

实现​

MemoryFilter 结构体​

// pkg/extproc/req_filter_memory.go

type MemoryFilter struct {
store memory.Store // Interface - can be MilvusStore or InMemoryStore
}

func NewMemoryFilter(store memory.Store) *MemoryFilter {
return &MemoryFilter{store: store}
}

说明: store 为第 8 节的 Store 接口,而非具体实现。运行时通常为生产环境 MilvusStore 或测试用 InMemoryStore。

记忆决策(复用现有流水线)​

已知限制: IsFact 分类器面向通识事实核查(如「法国首都是哪里?」)。可能将个人事实问题(如「我的预算是多少?」)误判为 fact,从而跳过记忆。

POC 缓解: 增加个人指代检测。若查询含人称代词("my", "I", "me"),则覆盖 IsFact,仍检索记忆。

未来: 重训或增强 fact 分类器以区分通识与个人事实。

// pkg/extproc/req_filter_memory.go

// shouldSearchMemory decides if query should trigger memory search
// Reuses existing pipeline classification signals with personal-fact override
func shouldSearchMemory(ctx *RequestContext, query string) bool {
// Check for personal indicators (overrides IsFact for personal questions)
hasPersonalIndicator := containsPersonalPronoun(query)

// 1. Fact query → skip UNLESS it contains personal pronouns
if ctx.IsFact && !hasPersonalIndicator {
logging.Debug("Memory: Skipping - general fact query")
return false
}

// 2. Tool required → skip (tool provides answer)
if ctx.RequiresTool {
logging.Debug("Memory: Skipping - tool query")
return false
}

// 3. Greeting/social → skip (no context needed)
if isGreeting(query) {
logging.Debug("Memory: Skipping - greeting")
return false
}

// 4. Default: search memory (conservative - don't miss context)
return true
}

func containsPersonalPronoun(query string) bool {
// Simple check for personal context indicators
personalPatterns := regexp.MustCompile(`(?i)\b(my|i|me|mine|i'm|i've|i'll)\b`)
return personalPatterns.MatchString(query)
}

func isGreeting(query string) bool {
// Match greetings that are ONLY greetings, not "Hi, what's my budget?"
lower := strings.ToLower(strings.TrimSpace(query))

// Short greetings only (< 20 chars and matches pattern)
if len(lower) > 20 {
return false
}

greetings := []string{
`^(hi|hello|hey|howdy)[\s\!\.\,]*$`,
`^(hi|hello|hey)[\s\,]*(there)?[\s\!\.\,]*$`,
`^(thanks|thank you|thx)[\s\!\.\,]*$`,
`^(bye|goodbye|see you)[\s\!\.\,]*$`,
`^(ok|okay|sure|yes|no)[\s\!\.\,]*$`,
}
for _, p := range greetings {
if regexp.MustCompile(p).MatchString(lower) {
return true
}
}
return false
}

上下文构建​

// buildSearchQuery builds an effective search query from history + current query
// MVP: context prepend, v1: LLM rewrite for vague queries
func buildSearchQuery(history []Message, query string) string {
// If query is self-contained, use as-is
if isSelfContained(query) {
return query
}

// MVP: Simple context prepend
context := summarizeHistory(history)
return query + " " + context

// v1 (future): LLM rewrite for vague queries
// if isVague(query) {
// return rewriteWithLLM(history, query)
// }
}

func isSelfContained(query string) bool {
// Self-contained: "What's my budget for the Hawaii trip?"
// NOT self-contained: "How much?", "And that one?", "What about it?"

vaguePatterns := []string{`^how much\??$`, `^what about`, `^and that`, `^this one`}
for _, p := range vaguePatterns {
if regexp.MustCompile(`(?i)`+p).MatchString(query) {
return false
}
}
return len(query) > 20 // Short queries are often vague
}

func summarizeHistory(history []Message) string {
// Extract key terms from last 3 user messages
var terms []string
count := 0
for i := len(history) - 1; i >= 0 && count < 3; i-- {
if history[i].Role == "user" {
terms = append(terms, extractKeyTerms(history[i].Content))
count++
}
}
return strings.Join(terms, " ")
}

// v1: LLM-based query rewriting (future enhancement)
func rewriteWithLLM(history []Message, query string) string {
prompt := fmt.Sprintf(`Conversation context: %s

Rewrite this vague query to be self-contained: "%s"
Return ONLY the rewritten query.`, summarizeHistory(history), query)

// Call LLM endpoint
resp, _ := http.Post(llmEndpoint+"/v1/chat/completions", ...)
return parseResponse(resp)
// "how much?" → "What is the budget for the Hawaii vacation?"
}

完整检索​

// pkg/extproc/req_filter_memory.go

func (f *MemoryFilter) RetrieveMemories(
ctx context.Context,
query string,
userID string,
history []Message,
) ([]*memory.RetrieveResult, error) {

// 1. Memory decision (skip if fact/tool/greeting)
if !shouldSearchMemory(ctx, query) {
logging.Debug("Memory: Skipping - not memory-relevant")
return nil, nil
}

// 2. Build search query (context prepend or LLM rewrite)
searchQuery := buildSearchQuery(history, query)

// 3. Search Milvus
results, err := f.store.Retrieve(ctx, memory.RetrieveOptions{
Query: searchQuery,
UserID: userID,
Limit: 5,
Threshold: 0.6,
})
if err != nil {
return nil, err
}

logging.Infof("Memory: Retrieved %d memories", len(results))
return results, nil
}

// InjectMemories adds memories to the LLM request
func (f *MemoryFilter) InjectMemories(
requestBody []byte,
memories []*memory.RetrieveResult,
) ([]byte, error) {
if len(memories) == 0 {
return requestBody, nil
}

// Format memories as context
var sb strings.Builder
sb.WriteString("## User's Relevant Context\n\n")
for _, mem := range memories {
sb.WriteString(fmt.Sprintf("- %s\n", mem.Memory.Content))
}

// Add as system message
return injectSystemMessage(requestBody, sb.String())
}

6. 记忆写入​

触发条件​

记忆提取由三类事件触发:

触发说明状态
每 N 轮每 10 轮提取一次MVP
会话结束会话结束时生成情景摘要未来
上下文漂移主题显著变化时提取未来

说明: 会话结束与漂移检测需额外实现。MVP 仅依赖「每 N 轮」触发。

流程​

┌─────────────────────────────────────────────────────────────────────────┐
│ MEMORY SAVING FLOW │
├─────────────────────────────────────────────────────────────────────────┤
│ │
│ TRIGGERS: │
│ ───────── │
│ ├── Every N turns (e.g., 10) ← MVP │
│ ├── End of session ← Future (needs detection) │
│ └── Context drift detected ← Future (needs detection) │
│ │
│ Runs: Async (background) - no user latency │
│ │
│ 1. GET BATCH │
│ ───────── │
│ Get last 10-15 turns from session │
│ │
│ 2. LLM EXTRACTION │
│ ────────────── │
│ Prompt: "Extract important facts. Include context. │
│ Return JSON: [{type, content}, ...]" │
│ │
│ LLM returns: │
│ [{"type": "semantic", "content": "budget for Hawaii is $10K"}] │
│ │
│ 3. DEDUPLICATION │
│ ───────────── │
│ For each extracted fact: │
│ - Embed content │
│ - Search existing memories (same user, same type) │
│ - If similarity > 0.9: UPDATE existing (merge/replace) │
│ - If similarity 0.7-0.9: CREATE new (gray zone, conservative) │
│ - If similarity < 0.7: CREATE new │
│ │
│ Example: │
│ Existing: "User's budget for Hawaii is $10,000" │
│ New: "User's budget is now $15,000" │
│ → Similarity ~0.92 → UPDATE existing with new value │
│ │
│ 4. STORE IN MILVUS │
│ ─────────────── │
│ Memory { id, type, content, embedding, user_id, created_at } │
│ │
│ 5. SESSION END (future): Create episodic summary │
│ ───────────────────────────────────────────── │
│ "On Dec 29, user planned Hawaii vacation with $10K budget" │
│ │
└─────────────────────────────────────────────────────────────────────────┘

关于 user_id: 此处指已登录用户(经认证的身份),而非当前会话中的匿名会话用户;具体映射需在 semantic router agent 侧配置。

实现​

// pkg/memory/extractor.go

type MemoryExtractor struct {
store memory.Store // Interface - can be MilvusStore or InMemoryStore
llmEndpoint string // LLM endpoint for fact extraction
batchSize int // Extract every N turns (default: 10)
turnCounts map[string]int
mu sync.Mutex
}

// ProcessResponse extracts and stores memories (runs async)
//
// Triggers (MVP: only first one implemented):
// - Every N turns (e.g., 10) ← MVP
// - End of session ← Future: needs session end detection
// - Context drift detected ← Future: needs drift detection
//
func (e *MemoryExtractor) ProcessResponse(
ctx context.Context,
sessionID string,
userID string,
history []Message,
) error {
e.mu.Lock()
e.turnCounts[sessionID]++
turnCount := e.turnCounts[sessionID]
e.mu.Unlock()

// MVP: Only extract every N turns
// Future: Also trigger on session end or context drift
if turnCount % e.batchSize != 0 {
return nil
}

// Get recent batch
batchStart := max(0, len(history) - e.batchSize - 5)
batch := history[batchStart:]

// LLM extraction
extracted, err := e.extractWithLLM(batch)
if err != nil {
return err
}

// Store with deduplication
for _, fact := range extracted {
existing, similarity := e.findSimilar(ctx, userID, fact.Content, fact.Type)

if similarity > 0.9 && existing != nil {
// Very similar → UPDATE existing memory
existing.Content = fact.Content // Use newer content
existing.UpdatedAt = time.Now()
if err := e.store.Update(ctx, existing.ID, existing); err != nil {
logging.Warnf("Failed to update memory: %v", err)
}
continue
}

// similarity < 0.9 → CREATE new memory
mem := &Memory{
ID: generateID("mem"),
Type: fact.Type,
Content: fact.Content,
UserID: userID,
Source: "conversation",
CreatedAt: time.Now(),
}

if err := e.store.Store(ctx, mem); err != nil {
logging.Warnf("Failed to store memory: %v", err)
}
}

return nil
}

// findSimilar searches for existing similar memories
func (e *MemoryExtractor) findSimilar(
ctx context.Context,
userID string,
content string,
memType MemoryType,
) (*Memory, float32) {
results, err := e.store.Retrieve(ctx, memory.RetrieveOptions{
Query: content,
UserID: userID,
Types: []MemoryType{memType},
Limit: 1,
Threshold: 0.7, // Only consider reasonably similar
})
if err != nil || len(results) == 0 {
return nil, 0
}
return results[0].Memory, results[0].Score
}

// extractWithLLM uses LLM to extract facts
//
// ⚠️ POC Limitation: LLM extraction is best-effort. Failures are logged but do not
// block the response. Incorrect extractions may occur.
//
// Future: Self-correcting memory (see Section 14 - Future Enhancements):
// - Track memory usage (access_count, last_accessed)
// - Score memories based on usage + age + retrieval feedback
// - Periodically prune low-score, unused memories
// - Detect contradictions → auto-merge or flag for resolution
//
func (e *MemoryExtractor) extractWithLLM(messages []Message) ([]ExtractedFact, error) {
prompt := `Extract important information from these messages.

IMPORTANT: Include CONTEXT for each fact.

For each piece of information:
- Type: "semantic" (facts, preferences) or "procedural" (instructions, how-to)
- Content: The fact WITH its context

BAD: {"type": "semantic", "content": "budget is $10,000"}
GOOD: {"type": "semantic", "content": "budget for Hawaii vacation is $10,000"}

Messages:
` + formatMessages(messages) + `

Return JSON array (empty if nothing to remember):
[{"type": "semantic|procedural", "content": "fact with context"}]`

// Call LLM with timeout
ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel()

reqBody := map[string]interface{}{
"model": "qwen3",
"messages": []map[string]string{
{"role": "user", "content": prompt},
},
}
jsonBody, _ := json.Marshal(reqBody)

req, _ := http.NewRequestWithContext(ctx, "POST",
e.llmEndpoint+"/v1/chat/completions",
bytes.NewReader(jsonBody))
req.Header.Set("Content-Type", "application/json")

resp, err := http.DefaultClient.Do(req)
if err != nil {
logging.Warnf("Memory extraction LLM call failed: %v", err)
return nil, err // Caller handles gracefully
}
defer resp.Body.Close()

if resp.StatusCode != 200 {
logging.Warnf("Memory extraction LLM returned %d", resp.StatusCode)
return nil, fmt.Errorf("LLM returned %d", resp.StatusCode)
}

facts, err := parseExtractedFacts(resp.Body)
if err != nil {
// JSON parse error - LLM returned malformed output
logging.Warnf("Memory extraction parse failed: %v", err)
return nil, err // Skip this batch, don't store garbage
}

return facts, nil
}

7. 记忆操作​

可对记忆执行的全部操作,由 Store 接口实现(见 第 8 节)。

操作说明触发接口方法状态
Store将新记忆写入 Milvus自动(LLM 提取)或显式 APIStore()MVP
Retrieve语义检索相关记忆查询时自动Retrieve()MVP
Update修改已有记忆内容去重或显式 APIUpdate()MVP
Forget按 ID 删除单条记忆显式 APIForget()MVP
ForgetByScope按用户/项目删除全部显式 APIForgetByScope()MVP
Consolidate合并相关记忆为摘要定时/达阈值Consolidate()未来
Reflect从记忆模式生成洞察智能体发起Reflect()未来

Forget 操作​

// Forget single memory
DELETE /v1/memory/{memory_id}

// Forget all memories for a user
DELETE /v1/memory?user_id=user_123

// Forget all memories for a project
DELETE /v1/memory?user_id=user_123&project_id=project_abc

用例:

  • 用户要求「忘掉关于 X 的内容」
  • GDPR/隐私合规(被遗忘权)
  • 清除过时信息

未来:Consolidate​

将多条相关记忆合并为一条摘要:

Before:
- "Budget for Hawaii is $10,000"
- "Added $2,000 to Hawaii budget"
- "Final Hawaii budget is $12,000"

After consolidation:
- "Hawaii trip budget: $12,000 (updated from initial $10,000)"

触发方式: 记忆条数超阈值、定时后台任务、会话结束。

未来:Reflect​

通过分析记忆模式生成洞察:

Input: All memories for user_123 about "deployment"

Output (Insight):
- "User frequently deploys payment-service (12 times)"
- "Common issue: port conflicts"
- "Preferred approach: docker-compose"

用例: 智能体可基于模式主动提供帮助。


8. 数据结构​

Memory​

// pkg/memory/types.go

type MemoryType string

const (
MemoryTypeEpisodic MemoryType = "episodic"
MemoryTypeSemantic MemoryType = "semantic"
MemoryTypeProcedural MemoryType = "procedural"
)

type Memory struct {
ID string `json:"id"`
Type MemoryType `json:"type"`
Content string `json:"content"`
Embedding []float32 `json:"-"`
UserID string `json:"user_id"`
ProjectID string `json:"project_id,omitempty"`
Source string `json:"source,omitempty"`
CreatedAt time.Time `json:"created_at"`
AccessCount int `json:"access_count"`
Importance float32 `json:"importance"`
}

Store 接口​

// pkg/memory/store.go

type Store interface {
// MVP Operations
Store(ctx context.Context, memory *Memory) error // Save new memory
Retrieve(ctx context.Context, opts RetrieveOptions) ([]*RetrieveResult, error) // Semantic search
Get(ctx context.Context, id string) (*Memory, error) // Get by ID
Update(ctx context.Context, id string, memory *Memory) error // Modify existing
Forget(ctx context.Context, id string) error // Delete by ID
ForgetByScope(ctx context.Context, scope MemoryScope) error // Delete by scope

// Utility
IsEnabled() bool
Close() error

// Future Operations (not yet implemented)
// Consolidate(ctx context.Context, memoryIDs []string) (*Memory, error) // Merge memories
// Reflect(ctx context.Context, scope MemoryScope) ([]*Insight, error) // Generate insights
}

9. API 扩展​

请求(已有)​

// pkg/responseapi/types.go

type ResponseAPIRequest struct {
// ... existing fields ...
MemoryConfig *MemoryConfig `json:"memory_config,omitempty"`
MemoryContext *MemoryContext `json:"memory_context,omitempty"`
}

type MemoryConfig struct {
Enabled bool `json:"enabled"`
MemoryTypes []string `json:"memory_types,omitempty"`
RetrievalLimit int `json:"retrieval_limit,omitempty"`
SimilarityThreshold float32 `json:"similarity_threshold,omitempty"`
AutoStore bool `json:"auto_store,omitempty"`
}

type MemoryContext struct {
UserID string `json:"user_id"`
ProjectID string `json:"project_id,omitempty"`
}

请求示例​

{
"model": "qwen3",
"input": "What's my budget for the trip?",
"previous_response_id": "resp_abc123",
"memory_config": {
"enabled": true,
"auto_store": true
},
"memory_context": {
"user_id": "user_456"
}
}

10. 配置​

# config.yaml
memory:
enabled: true
auto_store: true # Enable automatic fact extraction

milvus:
address: "milvus:19530"
collection: "agentic_memory"
dimension: 384 # Must match embedding model output

# Embedding model for memory
embedding:
model: "all-MiniLM-L6-v2" # 384-dim, optimized for semantic similarity
dimension: 384

# Retrieval settings
default_retrieval_limit: 5
default_similarity_threshold: 0.6 # Tunable; start conservative

# Extraction runs every N conversation turns
extraction_batch_size: 10

# External models for memory LLM features
# Query rewriting and fact extraction are enabled by adding external_models
external_models:
- llm_provider: "vllm"
model_role: "memory_rewrite" # Enables query rewriting
llm_endpoint:
address: "qwen"
port: 8000
llm_model_name: "qwen3"
llm_timeout_seconds: 30
max_tokens: 100
temperature: 0.1
- llm_provider: "vllm"
model_role: "memory_extraction" # Enables fact extraction
llm_endpoint:
address: "qwen"
port: 8000
llm_model_name: "qwen3"
llm_timeout_seconds: 30
max_tokens: 500
temperature: 0.1

配置说明​

参数取值理由
dimension: 384固定须与 all-MiniLM-L6-v2 输出一致
default_similarity_threshold: 0.6起始值依检索质量日志调参
extraction_batch_size: 10默认新鲜度与 LLM 成本平衡
llm_timeout_seconds: 30默认防止提取无限阻塞

嵌入模型选择:

模型维度优点缺点
all-MiniLM-L6-v2(POC 选用)384语义相似度更好、措辞容错高,适合记忆检索与去重需单独加载模型
Qwen3-Embedding-0.6B(已有)1024语义缓存已加载,无额外内存对措辞更敏感,可能漏检相近记忆

记忆为何用 384 维? 较低维度更能捕获高层语义,对具体数字、姓名等细节不敏感,有利于:

  • 检索:「我的预算是多少?」与「夏威夷行程预算是 $10K」即使措辞不同也能匹配
  • 去重:「预算 10K」与「预算现为10K」与「预算现为 15K」可识别为同一主题(更新数值)
  • 跨会话:不同会话间措辞自然不同

替代方案: 可复用 Qwen3-Embedding(1024 维)以避免加载第二模型;代价是匹配略严,可能增加假阴性。


11. 失败模式与回退(POC)​

本节明确记录各组件失败时的行为。POC 范围内优先优雅降级,而非复杂恢复。

失败检测行为日志
Milvus 不可用Store 初始化连接错误本会话关闭记忆功能ERROR: Milvus unavailable, memory disabled
Milvus 检索超时上下文 deadline 超时跳过记忆注入,无记忆继续WARN: Memory search timeout, skipping
嵌入生成失败candle-binding 报错本请求跳过记忆WARN: Embedding failed, skipping memory
LLM 提取失败HTTP 错误或超时跳过提取,不保存记忆WARN: Extraction failed, batch skipped
LLM 返回非法 JSON解析错误跳过提取,不保存记忆WARN: Extraction parse failed
无历史ctx.ConversationHistory 为空仅用查询检索(无上下文前缀)DEBUG: No history, query-only search
阈值过高返回 0 条不注入记忆DEBUG: No memories above threshold
阈值过低大量无关结果噪声上下文(POC 可接受)DEBUG: Retrieved N memories

优雅降级原则​

即使记忆失败,请求也必须成功。 记忆是增强能力,不是硬依赖。所有记忆操作均包在错误处理中:记录日志并继续。

// Example: Memory retrieval with fallback
memories, err := memoryFilter.RetrieveMemories(ctx, query, userID, history)
if err != nil {
logging.Warnf("Memory retrieval failed: %v", err)
memories = nil // Continue without memories
}
// Proceed with request (memories may be nil/empty)

12. 成功标准(POC)​

功能标准​

标准如何验证通过条件
跨会话检索会话 A 存事实,会话 B 查询事实被检索并注入
用户隔离用户 A 存事实,用户 B 查询用户 B 看不到用户 A 的事实
优雅降级停止 Milvus 后发请求请求成功(无记忆)
提取执行对话后查日志出现 Memory: Stored N facts

质量标准(POC 后测量)​

指标目标测量方式
检索相关性多数注入记忆相关人工抽查 50 条
提取准确性多数提取事实正确人工抽查 50 条
延迟影响P50 增加 <50ms开关记忆对比

POC 范围: 仅验证功能标准。质量指标在 POC 后测量,用于调阈值与改进提取提示词。


13. 实现计划​

Phase 1:检索​

任务文件
记忆决策(复用 Fact/Tool 信号)pkg/extproc/req_filter_memory.go
从历史构建上下文pkg/extproc/req_filter_memory.go
Milvus 检索 + 阈值过滤pkg/memory/milvus_store.go
向请求注入记忆pkg/extproc/req_filter_memory.go
接入请求阶段pkg/extproc/processor_req_body.go

Phase 2:写入​

任务文件
创建 MemoryExtractorpkg/memory/extractor.go
基于 LLM 的事实提取pkg/memory/extractor.go
去重逻辑pkg/memory/extractor.go
接入响应阶段(异步)pkg/extproc/processor_res_body.go

Phase 3:测试与调参​

任务说明
单元测试记忆决策、提取、检索
集成测试端到端流程
阈值调参依结果调整相似度阈值

14. 后续增强​

上下文压缩(高优先级)​

问题: Response API 当前向 LLM 发送全部对话历史。200 轮会话意味着每次请求数千 token——成本高且可能触顶上下文。

方案: 用两类输出替换旧消息:

输出用途存储替代
Facts长期记忆Milvus(第 6 节已有)
Current state会话上下文Redis旧消息

要点: 当前状态应为结构化(非散文摘要),便于知识图谱:

{"topic": "Hawaii vacation", "budget": "$10K", "decisions": ["fly direct"], "open": ["which hotel?"]}
┌─────────────────────────────────────────────────────────────────────────┐
│ CONTEXT COMPRESSION FLOW │
├─────────────────────────────────────────────────────────────────────────┤
│ │
│ BACKGROUND (every 10 turns): │
│ 1. Extract facts (reuse Section 6) → save to Milvus │
│ 2. Build current state (structured JSON) → save to Redis │
│ │
│ ON REQUEST (turn N): │
│ Context = [current state from Redis] ← replaces old messages │
│ + [raw last 5 turns] ← recent context │
│ + [relevant memories] ← cross-session (Milvus) │
│ │
└─────────────────────────────────────────────────────────────────────────┘

实现改动:

文件修改
pkg/responseapi/translator.go用当前状态 + 最近若干轮取代全量历史
pkg/responseapi/context_manager.go新建:管理当前状态
Redis 配置当前状态带 TTL 存储

LLM 收到的上下文(替代全量历史):

Context sent to LLM:
1. Current state (structured JSON from Redis) ~100 tokens
2. Last 5 raw messages ~400 tokens
3. Relevant memories from Milvus ~150 tokens
─────────────────────────────────────────────
Total: ~650 tokens (vs 10K for full history)

与智能体记忆的协同:

  • 第 6 节的事实提取在压缩过程中运行 → 写入 Milvus
  • 当前状态替代旧消息 → 减少 token
  • 结构化格式 → 便于未来接 KG

收益:

  • Token 可控(成本可预测)
  • 上下文质量更好(结构化状态优于全量历史)
  • KG-ready:结构化状态可直接映射图节点/边
  • 可支撑超长会话(1000+ 轮)

写入触发​

能力说明做法
会话结束检测会话结束时触发提取超时 / 显式信号 / API
上下文漂移检测主题显著变化时触发轮次间嵌入相似度

存储层​

能力说明优先级
Redis 热缓存Milvus 前的快速访问层高
TTL 与过期自动删除旧记忆(Redis 原生)高

高级能力​

能力说明优先级
自纠正记忆跟踪使用、按访问/年龄打分、自动剪枝低分记忆高
矛盾检测检测冲突事实、自动合并或标记高
按类型路由检索按 semantic/procedural/episodic 检索中
每用户配额限制存储上限中
图存储多跳查询的记忆关系按需
时序索引时序查询与衰减打分按需
并发处理同用户多会话加锁中

已知 POC 限制(明确推迟)​

限制影响为何可接受
无并发控制同用户多会话竞态POC 少见;生产修复
无记忆上限重度用户可能无限积累Phase 3 加配额
未测备份恢复Milvus 盘故障可能丢数据基础持久化可用;生产验证 HA
无智能更新更正可能产生重复最新优先;Forget API 可用
无对抗防御提示注入可能污染记忆POC 信任输入;后续过滤


附录​

附录 A:反思记忆​

状态: 未来扩展 — 不在本 POC 范围内。

基于过往交互的自省与经验教训。参见 Reflexion 论文。

存储内容:

  • 错误或不优回答带来的洞察
  • 对回答风格的偏好学习
  • 改善后续交互的模式

示例:

  • 「上次部署步骤有误——下次应先核对 k8s 版本」
  • 「用户偏好技术内容用要点列表而非长段落」
  • 「预算类问题应给分项,而非只给总额」

为何未来再做: 需要评估回答质量并生成自反思,依赖核心记忆基础设施之上。


附录 B:文件树​

pkg/
├── extproc/
│ ├── processor_req_body.go (EXTEND) Integrate retrieval
│ ├── processor_res_body.go (EXTEND) Integrate extraction
│ └── req_filter_memory.go (EXTEND) Pre-filter, retrieval, injection
│
├── memory/
│ ├── extractor.go (NEW) LLM-based fact extraction
│ ├── store.go (existing) Store interface
│ ├── milvus_store.go (existing) Milvus implementation
│ └── types.go (existing) Memory types
│
├── responseapi/
│ └── types.go (existing) MemoryConfig, MemoryContext
│
└── config/
└── config.go (EXTEND) Add extraction config

检索记忆时,「how much?」等模糊查询需要上下文才有效。本附录说明查询改写策略。

问题​

History: ["Planning Hawaii vacation", "Looking at hotels"]
Query: "How much?"
→ Direct search for "How much?" won't find "Hawaii budget is $10,000"

方案 1:上下文前缀(MVP)​

简单拼接,无 LLM 调用,延迟约 0ms。

func buildSearchQuery(history []Message, query string) string {
context := extractKeyTerms(history) // "Hawaii vacation planning"
return query + " " + context // "How much? Hawaii vacation planning"
}

优点: 快、简单
缺点: 可能混入无关词

方案 2:LLM 查询改写​

用 LLM 将查询改写为自包含问句,延迟约 100–200ms。

func rewriteQuery(history []Message, query string) string {
prompt := `Given conversation about: %s
Rewrite this query to be self-contained: "%s"
Return ONLY the rewritten query.`
return llm.Complete(fmt.Sprintf(prompt, summarize(history), query))
}
// "How much?" → "What is the budget for the Hawaii vacation?"

优点: 自然、嵌入匹配更好
缺点: LLM 延迟与成本

方案 3:HyDE(假设文档嵌入)​

生成假设答案,对其嵌入而非对查询嵌入。

HyDE 解决的问题:

Query: "What's the cost?" → embeds as QUESTION style
Stored: "Budget is $10,000" → embeds as STATEMENT style
Result: Low similarity (style mismatch)

With HyDE:
Query → LLM generates: "The cost is approximately $10,000"
This embeds as STATEMENT style → matches stored memory!
func hydeRewrite(query string, history []Message) string {
prompt := `Based on this conversation: %s
Write a short factual answer to: "%s"`
return llm.Complete(fmt.Sprintf(prompt, summarize(history), query))
}
// "How much?" → "The budget for the Hawaii trip is approximately $10,000"

优点: 检索质量最好(弥合问句/文档风格差异)
缺点: 延迟最高(~200ms)、LLM 成本

建议​

阶段方案适用
MVP上下文前缀默认全部查询
v1LLM 改写模糊查询("how much?", "and that?")
v2HyDE观测到问句风格检索分数低之后

说明: HyDE 是依据线上表现做的优化,不是预测。当你发现相关记忆存在但检索不到时再上。

参考文献​

Query Rewriting:

  1. HyDE - Precise Zero-Shot Dense Retrieval without Relevance Labels (Gao et al., 2022) - Style bridging (question → document style)
  2. RRR - Query Rewriting for Retrieval-Augmented LLMs (Ma et al., 2023) - Trainable rewriter with RL, handles conversational context

Agentic Memory (from Issue #808):

  1. MemGPT - Towards LLMs as Operating Systems (Packer et al., 2023)
  2. Generative Agents - Simulacra of Human Behavior (Park et al., 2023)
  3. Reflexion - Language Agents with Verbal Reinforcement Learning (Shinn et al., 2023)
  4. Voyager - An Open-Ended Embodied Agent with LLMs (Wang et al., 2023)

文档作者:[Yehudit Kerido, Marina Koushnir]
最后更新:2025 年 12 月
状态:POC 设计 - v3(已评审修订)
基于:Issue #808 - Explore Agentic Memory in Response API