align add/search API with server product routes

README.md

@@ -5,8 +5,8 @@ Official plugin maintained by MemTensor.
 A minimal OpenClaw lifecycle plugin that **recalls** memories from MemOS Cloud before each run and **adds** new messages to MemOS Cloud after each run.
 
 ## Features
-- **Recall**: `before_agent_start` → `/search/memory`
-- **Add**: `agent_end` → `/add/message`
+- **Recall**: `before_agent_start` → `/product/search`
+- **Add**: `agent_end` → `/product/add`
 - Uses **Token** auth (`Authorization: Token <MEMOS_API_KEY>`)
 
 ## Install
@@ -92,7 +92,7 @@ MEMOS_API_KEY=YOUR_TOKEN
 - `MEMOS_API_KEY` (required; Token auth) — get it at https://memos-dashboard.openmem.net/cn/apikeys/
 - `MEMOS_USER_ID` (optional; default: `openclaw-user`)
 - `MEMOS_CONVERSATION_ID` (optional override)
-- `MEMOS_RECALL_GLOBAL` (default: `true`; when true, search does **not** pass conversation_id)
+- `MEMOS_RECALL_GLOBAL` (default: `true`; when true, search does **not** pass `session_id`)
 - `MEMOS_MULTI_AGENT_MODE` (default: `false`; enable multi-agent data isolation)
 - `MEMOS_CONVERSATION_PREFIX` / `MEMOS_CONVERSATION_SUFFIX` (optional)
 - `MEMOS_CONVERSATION_SUFFIX_MODE` (`none` | `counter`, default: `none`)
@@ -158,24 +158,24 @@ In `plugins.entries.memos-cloud-openclaw-plugin.config`:
 
 ## How it Works
 - **Recall** (`before_agent_start`)
-  - Builds a `/search/memory` request using `user_id`, `query` (= prompt + optional prefix), and optional filters.
-  - Default **global recall**: when `recallGlobal=true`, it does **not** pass `conversation_id`.
+  - Builds a `/product/search` request using `user_id`, `query` (= prompt + optional prefix), and optional filters.
+  - Default **global recall**: when `recallGlobal=true`, it does **not** pass `session_id`.
   - Optional second-pass filtering: if `recallFilterEnabled=true`, candidates are sent to your configured model and only returned `keep` items are injected.
   - Injects a stable MemOS recall protocol via `appendSystemContext`, while the retrieved `<memories>` block remains in `prependContext`.
 
 - **Add** (`agent_end`)
-  - Builds a `/add/message` request with the **last turn** by default (user + assistant).
-  - Sends `messages` with `user_id`, `conversation_id`, and optional `tags/info/agent_id/app_id`.
+  - Builds a `/product/add` request with the **last turn** by default (user + assistant).
+  - Sends `messages` with `user_id`, `session_id`, and optional `tags/info/agent_id/app_id`.
 
 ## Multi-Agent Support
 The plugin provides native support for multi-agent architectures (via the `agent_id` parameter):
 - **Enable Mode**: Set `"multiAgentMode": true` in config or `MEMOS_MULTI_AGENT_MODE=true` in env variables (default is `false`).
 - **Dynamic Context**: When enabled, it automatically captures `ctx.agentId` during OpenClaw lifecycle hooks. (Note: the default OpenClaw agent `"main"` is ignored to preserve backwards compatibility for single-agent users).
-- **Data Isolation**: The `agent_id` is automatically injected into both `/search/memory` and `/add/message` requests. This ensures completely isolated memory and message histories for different agents, even under the same user or session.
+- **Data Isolation**: The `agent_id` is automatically injected into both `/product/search` and `/product/add` requests. This ensures completely isolated memory and message histories for different agents, even under the same user or session.
 - **Static Override**: You can also force a specific agent ID by setting `"agentId": "your_agent_id"` in the plugin's `config`.
 
 ## Notes
-- `conversation_id` defaults to OpenClaw `sessionKey` (unless `conversationId` is provided). **TODO**: consider binding to OpenClaw `sessionId` directly.
+- `session_id` defaults to OpenClaw `sessionKey` (unless `conversationId` is provided). **TODO**: consider binding to OpenClaw `sessionId` directly.
 - Optional **prefix/suffix** via env or config; `conversationSuffixMode=counter` increments on `/new` (requires `hooks.internal.enabled`).
 
 ## Acknowledgements

README_ZH.md

@@ -7,8 +7,8 @@
 - **添加记忆**：在每轮对话结束后把消息写回 MemOS Cloud
 
 ## 功能
-- **Recall**：`before_agent_start` → `/search/memory`
-- **Add**：`agent_end` → `/add/message`
+- **Recall**：`before_agent_start` → `/product/search`
+- **Add**：`agent_end` → `/product/add`
 - 使用 **Token** 认证（`Authorization: Token <MEMOS_API_KEY>`）
 
 ## 安装
@@ -94,7 +94,7 @@ MEMOS_API_KEY=YOUR_TOKEN
 - `MEMOS_API_KEY`（必填，Token 认证）—— 获取地址：https://memos-dashboard.openmem.net/cn/apikeys/
 - `MEMOS_USER_ID`（可选，默认 `openclaw-user`）
 - `MEMOS_CONVERSATION_ID`（可选覆盖）
-- `MEMOS_RECALL_GLOBAL`（默认 `true`；为 true 时检索不传 conversation_id）
+- `MEMOS_RECALL_GLOBAL`（默认 `true`；为 true 时检索不传 `session_id`）
 - `MEMOS_MULTI_AGENT_MODE`（默认 `false`；是否开启多 Agent 数据隔离模式）
 - `MEMOS_CONVERSATION_PREFIX` / `MEMOS_CONVERSATION_SUFFIX`（可选）
 - `MEMOS_CONVERSATION_SUFFIX_MODE`（`none` | `counter`，默认 `none`）
@@ -158,29 +158,29 @@ MEMOS_API_KEY=YOUR_TOKEN
 
 ## 工作原理
 ### 1) 召回（before_agent_start）
-- 组装 `/search/memory` 请求
+- 组装 `/product/search` 请求
   - `user_id`、`query`（= prompt + 可选前缀）
-  - 默认**全局召回**：`recallGlobal=true` 时不传 `conversation_id`
+  - 默认**全局召回**：`recallGlobal=true` 时不传 `session_id`
   - 可选 `filter` / `knowledgebase_ids`
 - （可选）若开启 `recallFilterEnabled`，会先把 `memory/preference/tool_memory` 候选发给你配置的模型做二次筛选，只保留 `keep` 的条目
 - 将稳定的 MemOS 召回协议通过 `appendSystemContext` 注入，而检索到的 `<memories>` 数据块继续通过 `prependContext` 注入
 
 ### 2) 添加（agent_end）
 - 默认只写**最后一轮**（user + assistant）
-- 构造 `/add/message` 请求：
-  - `user_id`、`conversation_id`
+- 构造 `/product/add` 请求：
+  - `user_id`、`session_id`
   - `messages` 列表
   - 可选 `tags / info / agent_id / app_id`
 
 ## 多Agent支持（Multi-Agent）
 插件内置对多Agent模式的支持（`agent_id` 参数）：
 - **开启模式**：需要在配置中设置 `"multiAgentMode": true` 或在环境变量中设置 `MEMOS_MULTI_AGENT_MODE=true`（默认为 `false`）。
 - **动态获取**：开启后，执行生命周期钩子时会自动读取上下文中的 `ctx.agentId`。（注：OpenClaw 的默认 Agent `"main"` 会被自动忽略，以保证老用户的单 Agent 数据兼容性）。
-- **数据隔离**：在调用 `/search/memory`（检索记忆）和 `/add/message`（添加记录）时会自动附带该 `agent_id`，从而保证即使是同一用户下的不同 Agent 之间，记忆和反馈数据也是完全隔离的。
+- **数据隔离**：在调用 `/product/search`（检索记忆）和 `/product/add`（添加记录）时会自动附带该 `agent_id`，从而保证即使是同一用户下的不同 Agent 之间，记忆和反馈数据也是完全隔离的。
 - **静态配置**：如果需要，也可在上述插件的 `config` 中显式指定 `"agentId": "your_agent_id"` 作为固定值。
 
 ## 说明
-- 未显式指定 `conversation_id` 时，默认使用 OpenClaw `sessionKey`。**TODO**：后续考虑直接绑定 OpenClaw `sessionId`。
+- 未显式指定 `session_id` 时，默认使用 OpenClaw `sessionKey`。**TODO**：后续考虑直接绑定 OpenClaw `sessionId`。
 - 可配置前后缀；`conversationSuffixMode=counter` 时会在 `/new` 递增（需 `hooks.internal.enabled`）。
 
 ## 致谢

clawdbot.plugin.json

@@ -23,15 +23,15 @@
       },
       "conversationId": {
         "type": "string",
-        "description": "Override conversation_id"
+        "description": "Override session_id"
       },
       "conversationIdPrefix": {
         "type": "string",
-        "description": "conversation_id prefix"
+        "description": "session_id prefix"
       },
       "conversationIdSuffix": {
         "type": "string",
-        "description": "conversation_id suffix"
+        "description": "session_id suffix"
       },
       "conversationSuffixMode": {
         "type": "string",

index.js

@@ -81,7 +81,7 @@ function buildSearchPayload(cfg, prompt, ctx) {
 
   if (!cfg.recallGlobal) {
     const conversationId = resolveConversationId(cfg, ctx);
-    if (conversationId) payload.conversation_id = conversationId;
+    if (conversationId) payload.session_id = conversationId;
   }
 
   let filterObj = cfg.filter ? JSON.parse(JSON.stringify(cfg.filter)) : null;
@@ -101,42 +101,41 @@ function buildSearchPayload(cfg, prompt, ctx) {
 
   if (filterObj) payload.filter = filterObj;
 
-  if (cfg.knowledgebaseIds?.length) payload.knowledgebase_ids = cfg.knowledgebaseIds;
+  if (cfg.knowledgebaseIds?.length) payload.readable_cube_ids = cfg.knowledgebaseIds;
 
-  payload.memory_limit_number = cfg.memoryLimitNumber;
+  payload.top_k = cfg.memoryLimitNumber;
   payload.include_preference = cfg.includePreference;
-  payload.preference_limit_number = cfg.preferenceLimitNumber;
-  payload.include_tool_memory = cfg.includeToolMemory;
-  payload.tool_memory_limit_number = cfg.toolMemoryLimitNumber;
+  payload.pref_top_k = cfg.preferenceLimitNumber;
+  payload.search_tool_memory = cfg.includeToolMemory;
+  payload.tool_mem_top_k = cfg.toolMemoryLimitNumber;
   payload.relativity = cfg.relativity;
 
   return payload;
 }
 
 function buildAddMessagePayload(cfg, messages, ctx) {
-  const payload = {
-    user_id: cfg.userId,
-    conversation_id: resolveConversationId(cfg, ctx),
-    messages,
-    source: MEMOS_SOURCE,
-  };
-
-  const agentId = getEffectiveAgentId(cfg, ctx);
-  if (agentId) payload.agent_id = agentId;
-  if (cfg.appId) payload.app_id = cfg.appId;
-  if (cfg.tags?.length) payload.tags = cfg.tags;
-
   const info = {
     source: "openclaw",
     sessionKey: ctx?.sessionKey,
     agentId: ctx?.agentId,
     ...(cfg.info || {}),
   };
-  if (Object.keys(info).length > 0) payload.info = info;
+  const agentId = getEffectiveAgentId(cfg, ctx);
+  if (agentId) info.agent_id = agentId;
+  if (cfg.appId) info.app_id = cfg.appId;
+  if (cfg.allowPublic !== undefined) info.allow_public = cfg.allowPublic;
+
+  const payload = {
+    user_id: cfg.userId,
+    session_id: resolveConversationId(cfg, ctx),
+    messages,
+    source: MEMOS_SOURCE,
+    info,
+    async_mode: cfg.asyncMode === "sync" || cfg.asyncMode === false ? "sync" : "async",
+  };
 
-  payload.allow_public = cfg.allowPublic;
-  if (cfg.allowKnowledgebaseIds?.length) payload.allow_knowledgebase_ids = cfg.allowKnowledgebaseIds;
-  payload.async_mode = cfg.asyncMode;
+  if (cfg.tags?.length) payload.custom_tags = cfg.tags;
+  if (cfg.allowKnowledgebaseIds?.length) payload.writable_cube_ids = cfg.allowKnowledgebaseIds;
 
   return payload;
 }

lib/memos-cloud-api.js

@@ -60,7 +60,56 @@ function stripQuotes(value) {
 
 export function extractResultData(result) {
   if (!result || typeof result !== "object") return null;
-  return result.data ?? result.data?.data ?? result.data?.result ?? null;
+  const data = result.data ?? result.data?.data ?? result.data?.result ?? null;
+  return normalizeSearchData(data);
+}
+
+function flattenSearchBucket(buckets, valueMapper) {
+  if (!Array.isArray(buckets)) return [];
+  const list = [];
+  for (const bucket of buckets) {
+    const memories = Array.isArray(bucket?.memories) ? bucket.memories : [];
+    for (const item of memories) {
+      const value = valueMapper(item);
+      if (!value) continue;
+      list.push({
+        ...item,
+        ...value,
+        relativity: item?.metadata?.relativity ?? item?.metadata?.score ?? 1,
+        create_time: item?.metadata?.create_time ?? item?.metadata?.created_at ?? item?.created_at,
+      });
+    }
+  }
+  return list;
+}
+
+function normalizeSearchData(data) {
+  if (!data || typeof data !== "object") return data;
+  const hasLegacyShape =
+    Array.isArray(data.memory_detail_list) ||
+    Array.isArray(data.preference_detail_list) ||
+    Array.isArray(data.tool_memory_detail_list);
+  if (hasLegacyShape) return data;
+
+  const hasServerShape =
+    Array.isArray(data.text_mem) || Array.isArray(data.pref_mem) || Array.isArray(data.tool_mem);
+  if (!hasServerShape) return data;
+
+  return {
+    ...data,
+    memory_detail_list: flattenSearchBucket(data.text_mem, (item) => ({
+      memory_value: item?.memory || item?.metadata?.memory || "",
+      memory_key: item?.memory || item?.metadata?.memory || "",
+    })),
+    preference_detail_list: flattenSearchBucket(data.pref_mem, (item) => ({
+      preference: item?.memory || item?.metadata?.memory || "",
+      preference_type: item?.metadata?.preference_type || "",
+    })),
+    tool_memory_detail_list: flattenSearchBucket(data.tool_mem, (item) => ({
+      tool_value: item?.memory || item?.metadata?.memory || "",
+    })),
+    preference_note: data.pref_note ?? data.preference_note ?? "",
+  };
 }
 
 function pad2(value) {
@@ -318,7 +367,7 @@ function sanitizeAddMessageEntry(entry) {
 }
 
 export async function searchMemory(cfg, payload) {
-  return callApi(cfg, "/search/memory", sanitizeSearchPayload(payload));
+  return callApi(cfg, "/product/search", payload);
 }
 
 export async function addMessage(cfg, payload) {
@@ -329,7 +378,7 @@ export async function addMessage(cfg, payload) {
     // Fail open: if sanitization throws unexpectedly, send original payload.
     finalPayload = payload;
   }
-  return callApi(cfg, "/add/message", finalPayload);
+  return callApi(cfg, "/product/add", finalPayload);
 }
 
 function isInboundMetaSentinelLine(line) {

moltbot.plugin.json

@@ -23,15 +23,15 @@
       },
       "conversationId": {
         "type": "string",
-        "description": "Override conversation_id"
+        "description": "Override session_id"
       },
       "conversationIdPrefix": {
         "type": "string",
-        "description": "conversation_id prefix"
+        "description": "session_id prefix"
       },
       "conversationIdSuffix": {
         "type": "string",
-        "description": "conversation_id suffix"
+        "description": "session_id suffix"
       },
       "conversationSuffixMode": {
         "type": "string",

openclaw.plugin.json

@@ -23,15 +23,15 @@
       },
       "conversationId": {
         "type": "string",
-        "description": "Override conversation_id"
+        "description": "Override session_id"
       },
       "conversationIdPrefix": {
         "type": "string",
-        "description": "conversation_id prefix"
+        "description": "session_id prefix"
       },
       "conversationIdSuffix": {
         "type": "string",
-        "description": "conversation_id suffix"
+        "description": "session_id suffix"
       },
       "conversationSuffixMode": {
         "type": "string",