feishu stream (#6629)

86758af5 · Archer · GitHub · 25d6bb7d · 86758af5 · 86758af5
Commit 86758af5 authored Mar 24, 2026 by Archer Committed by GitHub Mar 24, 2026
7 changed files
--- a/.claude/design/wechat-file-support.md
+++ b/.claude/design/wechat-file-support.md
+# 微信个人号文件接收方案设计
+> 文件路径：`packages/service/support/outLink/wechat/`
+## 一、背景与现状
+### 当前架构
+```
+ILink API (长轮询) → mq.ts → messageParser.ts → outlinkInvokeChat()
+```
+- **消息拉取**：`ILinkClient.getUpdates()` 长轮询，返回 `WeixinMessage[]`
+- **消息解析**：`messageParser.ts` 只处理 `text`（type=1）和 `voice`（type=3）
+- **查询构造**：仅传递 `{ text: { content: group.text } }` 给 workflow
+- **文件支持**：❌ 完全缺失
+### FastGPT 已有基础设施
+| 能力 | 实现位置 | 说明 |
+|------|----------|------|
+| 图片存储 | `file/image/controller.ts` → `uploadMongoImg()` | base64 存 MongoDB，返回内部 URL |
+| 文件存储 | `file/gridfs/` (BucketNameEnum.chat) | GridFS 流式存储 |
+| Chat 输入类型 | `global/core/chat/type.ts` → `UserChatItemValueItemType` | 支持 `text` + `file(image/file)` |
+### `UserChatItemValueItemType` 结构
+```typescript
+type UserChatItemValueItemType = {
+  text?: { content: string };
+  file?: {
+    type: 'image' | 'file';   // ChatFileTypeEnum
+    name?: string;
+    key?: string;
+    url: string;              // 内部访问 URL
+  };
+};
+```
+---
+## 二、目标
+支持微信用户通过 iLink 渠道向 FastGPT 发送：
+| 类型 | 来源 | 处理方式 |
+|------|------|----------|
+| 图片 | `item_list[].type = 2` | 下载 → 存 MongoImage → 内部 URL |
+| 文件/附件 | `item_list[].type = 5` | 下载 → 存 GridFS(chat) → 内部 URL |
+| 视频 | `item_list[].type = 4` | 下载 → 存 GridFS(chat) → 内部 URL（可选）|
+---
+## 三、S3 存储方案
+### 核心选择：`S3ChatSource.uploadChatFileByBuffer()`
+| 字段 | 值 |
+|------|----|
+| 存储路径 | `chat/{appId}/{uId}/{chatId}/{filename}` |
+| 存储桶 | Private Bucket（私有，需鉴权访问） |
+| TTL | 1 小时（由 `MongoS3TTL` 自动清理） |
+| 访问 URL | `createGetChatFileURL({ key, expiredHours: 24 })` 生成预签名 URL |
+### 调用关系
+```
+getS3ChatSource()
+  └── uploadChatFileByBuffer({ appId, chatId, uId, filename, buffer, contentType })
+        └── getFileS3Key.chat(...)  →  key = "chat/{appId}/{uId}/{chatId}/{filename}"
+        └── uploadFileByBuffer({ key, buffer, contentType })
+              ├── MongoS3TTL.create({ expiredTime: +1h })   // 文件 1 小时后自动清理
+              ├── client.uploadObject(...)
+              └── returns { key, accessUrl }               // accessUrl 预签名 2h
+```
+### URL 传入 workflow
+```typescript
+// UserChatItemValueItemType.file
+{
+  type: 'image' | 'file',   // ChatFileTypeEnum
+  name: string,             // 原始文件名
+  key: string,              // S3 object key，用于后续删除/引用
+  url: string               // 预签名 URL（accessUrl，2h 内有效）
+}
+```
+> 由于 workflow 在文件上传后立即执行，2 小时访问窗口完全够用。
+---
+## 四、ILink API 消息结构（推断）
+基于 iLink WeChat Bot API 常见规范，图片/文件消息的 `item_list` 元素结构：
+```typescript
+// 图片消息项（type = 2）
+type ImageMessageItem = {
+  type: 2;
+  image_item?: {
+    cdn_url: string;       // 图片 CDN 直链（有效期有限）
+    aes_key?: string;      // 加密 key（若内容加密）
+    file_size?: number;    // 字节数
+    width?: number;
+    height?: number;
+  };
+};
+// 文件消息项（type = 5）
+type FileMessageItem = {
+  type: 5;
+  file_item?: {
+    file_id?: string;      // 文件 ID（用于获取下载链接）
+    file_name: string;     // 原始文件名
+    cdn_url?: string;      // 文件 CDN 直链（有效期有限）
+    file_size?: number;
+    file_ext?: string;     // 扩展名 e.g. "pdf"
+  };
+};
+```
+> **注意**：实际字段需与 iLink API 文档对齐。本方案设计为可扩展结构，字段名通过常量隔离，便于修正。
+---
+## 四、方案设计
+### 4.1 整体流程
+```
+ILink getUpdates()
+   ↓
+MessageItem[]
+   ↓ (messageParser)
+ParsedMessageGroup {
+  userId, text, contextToken, msgIds,
+  files: [{ type, tempUrl, name }]  ← 新增
+}
+   ↓ (mq.ts - processUserGroup)
+downloadAndStoreFiles()   ← 新增
+   ↓
+UserChatItemValueItemType[] = [
+  { text: { content: "..." } },
+  { file: { type: "image", url: "http://..." } },
+  { file: { type: "file", url: "http://...", name: "report.pdf" } }
+]
+   ↓
+outlinkInvokeChat({ query: [...] })
+```
+### 4.2 变更范围
+共涉及 **4 个文件**：
+| 文件 | 变更类型 | 说明 |
+|------|----------|------|
+| `ilinkClient.ts` | 扩展类型 + 新增方法 | 添加图片/文件 Item 类型；添加 `downloadMedia()` |
+| `messageParser.ts` | 扩展逻辑 + 类型 | 解析图片/文件 item，收集 `tempUrl` |
+| `wechat/fileHandler.ts` | **新建** | 下载 → 存储 → 返回内部 URL |
+| `mq.ts` | 扩展逻辑 | 调用 fileHandler，拼装 query 数组 |
+---
+## 五、详细设计
+### 5.1 `ilinkClient.ts` — 扩展类型与下载能力
+#### 新增类型
+```typescript
+// 在 MessageItem 基础上扩展
+export type MessageItem = {
+  type: number;
+  text_item?: { text: string };
+  voice_item?: { text: string };
+  ref_msg?: { title?: string };
+  // 新增
+  image_item?: {
+    cdn_url?: string;
+    file_size?: number;
+    width?: number;
+    height?: number;
+  };
+  file_item?: {
+    file_id?: string;
+    file_name?: string;
+    cdn_url?: string;
+    file_size?: number;
+    file_ext?: string;
+  };
+};
+// 媒体下载结果
+export type MediaDownloadResult = {
+  buffer: Buffer;
+  mimeType: string;
+  fileName?: string;
+  fileSize: number;
+};
+```
+#### 新增方法 `downloadMedia()`
+```typescript
+// 最大允许下载文件大小（20MB）
+const MAX_MEDIA_SIZE = 20 * 1024 * 1024;
+const DOWNLOAD_TIMEOUT_MS = 30_000;
+async downloadMedia(url: string, fileName?: string): Promise<MediaDownloadResult> {
+  const controller = new AbortController();
+  const timer = setTimeout(() => controller.abort(), DOWNLOAD_TIMEOUT_MS);
+  try {
+    const res = await fetch(url, {
+      headers: this.buildHeaders(),  // 携带 token 鉴权
+      signal: controller.signal
+    });
+    clearTimeout(timer);
+    if (!res.ok) throw new Error(`Media download failed: HTTP ${res.status}`);
+    const contentLength = Number(res.headers.get('content-length') || 0);
+    if (contentLength > MAX_MEDIA_SIZE) {
+      throw new Error(`File too large: ${contentLength} bytes`);
+    }
+    const arrayBuffer = await res.arrayBuffer();
+    const buffer = Buffer.from(arrayBuffer);
+    if (buffer.length > MAX_MEDIA_SIZE) {
+      throw new Error(`File too large: ${buffer.length} bytes`);
+    }
+    const contentType = res.headers.get('content-type') || 'application/octet-stream';
+    const mimeType = contentType.split(';')[0].trim();
+    return { buffer, mimeType, fileName, fileSize: buffer.length };
+  } catch (err) {
+    clearTimeout(timer);
+    throw err;
+  }
+}
+```
+---
+### 5.2 `messageParser.ts` — 扩展解析逻辑
+#### 新增常量与类型
+```typescript
+const MSG_ITEM_IMAGE = 2;
+const MSG_ITEM_FILE = 5;
+// 解析出的媒体引用（尚未下载，只存 URL 和元信息）
+export type ParsedMediaRef = {
+  type: 'image' | 'file';
+  tempUrl: string;       // iLink CDN URL（有效期有限）
+  name?: string;         // 原始文件名
+};
+// 扩展 ParsedMessageGroup
+export type ParsedMessageGroup = {
+  userId: string;
+  text: string;
+  contextToken: string;
+  msgIds: string[];
+  mediaRefs: ParsedMediaRef[];   // 新增：待下载的媒体引用
+};
+```
+#### 更新 `extractTextFromItem` 和 `groupMessagesByUser`
+```typescript
+// extractTextFromItem 保持不变，仍只返回文本
+// 新增：从 item 提取媒体引用
+export function extractMediaRefFromItem(
+  item: NonNullable<WeixinMessage['item_list']>[number]
+): ParsedMediaRef | null {
+  if (item.type === MSG_ITEM_IMAGE && item.image_item?.cdn_url) {
+    return {
+      type: 'image',
+      tempUrl: item.image_item.cdn_url,
+      name: undefined
+    };
+  }
+  if (item.type === MSG_ITEM_FILE && item.file_item?.cdn_url) {
+    return {
+      type: 'file',
+      tempUrl: item.file_item.cdn_url,
+      name: item.file_item.file_name
+    };
+  }
+  return null;
+}
+```
+在 `groupMessagesByUser` 中：
+```typescript
+// 处理逻辑变更：消息中若包含文件但无文本，也纳入分组（原逻辑会跳过无文本消息）
+for (const msg of msgs) {
+  if (msg.message_type !== MSG_TYPE_USER) continue;
+  let text = '';
+  const mediaRefs: ParsedMediaRef[] = [];
+  for (const item of msg.item_list ?? []) {
+    if (!text) {
+      const t = extractTextFromItem(item);
+      if (t) text = t;
+    }
+    const ref = extractMediaRefFromItem(item);
+    if (ref) mediaRefs.push(ref);
+  }
+  // 关键变更：无文本但有媒体时也处理
+  if (!text && mediaRefs.length === 0) continue;
+  // ... 合并到分组（mediaRefs 追加合并）
+}
+```
+---
+### 5.3 `wechat/fileHandler.ts` — 新建文件处理模块
+图片和文件**统一存入 S3 私有桶**（`chat` source），通过 `S3ChatSource.uploadChatFileByBuffer()` 写入，
+上传完成后用 `createGetChatFileURL()` 生成预签名 URL 传给 workflow。
+```typescript
+import type { MediaDownloadResult } from './ilinkClient';
+import { getS3ChatSource } from '../../../common/s3/sources/chat';
+import type { UserChatItemFileItemType } from '@fastgpt/global/core/chat/type';
+import { ChatFileTypeEnum } from '@fastgpt/global/core/chat/constants';
+import mime from 'mime-types';
+export type MediaStoreResult = UserChatItemFileItemType;
+/**
+ * 统一入口：下载 iLink CDN 文件 → 存入 S3(chat) → 返回预签名 URL
+ *
+ * S3 key:  chat/{appId}/{uId}/{chatId}/{filename}
+ * TTL:     文件 1h（MongoS3TTL 自动清理）；预签名 URL 24h
+ */
+export async function downloadAndStoreMedia(params: {
+  type: 'image' | 'file';
+  tempUrl: string;
+  name?: string;
+  appId: string;
+  chatId: string;
+  uId: string;
+  downloadFn: (url: string, name?: string) => Promise<MediaDownloadResult>;
+}): Promise<MediaStoreResult | null> {
+  const { type, tempUrl, name, appId, chatId, uId, downloadFn } = params;
+  try {
+    const media = await downloadFn(tempUrl, name);
+    const ext = mime.extension(media.mimeType) || 'bin';
+    const filename = media.fileName || `wechat_${type}_${Date.now()}.${ext}`;
+    const chatSource = getS3ChatSource();
+    // 上传到 S3，返回 { key, accessUrl }
+    // accessUrl 是 createExternalUrl({ key, expiredHours: 2 }) 的预签名 URL
+    const { key, accessUrl } = await chatSource.uploadChatFileByBuffer({
+      appId,
+      chatId,
+      uId,
+      filename,
+      buffer: media.buffer,
+      contentType: media.mimeType
+    });
+    // 生成更长有效期的预签名 URL 供 workflow 访问（24h）
+    const url = await chatSource.createGetChatFileURL({
+      key,
+      expiredHours: 24,
+      external: false
+    });
+    return {
+      type: type === 'image' ? ChatFileTypeEnum.image : ChatFileTypeEnum.file,
+      name: filename,
+      key,
+      url
+    };
+  } catch (error) {
+    // 单个文件失败不中断整体流程，上层记录日志
+    return null;
+  }
+}
+```
+---
+### 5.4 `mq.ts` — 扩展 processUserGroup
+```typescript
+async function processUserGroup(
+  outLink: OutLinkSchema<WechatAppType>,
+  group: ParsedMessageGroup
+): Promise<void> {
+  const app = outLink.app;
+  const chatId = `wechat_${outLink.shareId}_${group.userId}`;
+  const client = new ILinkClient(app.baseUrl, app.token);
+  // 1. 并发下载所有媒体文件（最多 5 个，防止超载）
+  const maxFiles = 5;
+  const mediaRefs = group.mediaRefs.slice(0, maxFiles);
+  const fileResults = await Promise.allSettled(
+    mediaRefs.map((ref) =>
+      downloadAndStoreMedia({
+        type: ref.type,
+        tempUrl: ref.tempUrl,
+        name: ref.name,
+        teamId: outLink.teamId,
+        shareId: outLink.shareId,
+        downloadFn: (url, name) => client.downloadMedia(url, name)
+      })
+    )
+  );
+  // 2. 构造 UserChatItemValueItemType[]
+  const query: UserChatItemValueItemType[] = [];
+  // 文本（如有）
+  if (group.text) {
+    query.push({ text: { content: group.text } });
+  }
+  // 成功上传的文件
+  for (const result of fileResults) {
+    if (result.status === 'fulfilled' && result.value) {
+      query.push({ file: result.value });
+    }
+  }
+  // 若 query 为空（全部失败），回复提示
+  if (query.length === 0) {
+    await client.sendMessage({
+      to_user_id: group.userId,
+      text: '文件处理失败，请重试。',
+      context_token: group.contextToken
+    });
+    return;
+  }
+  try {
+    await outlinkInvokeChat({
+      outLinkConfig: outLink,
+      chatId,
+      query,
+      messageId: group.msgIds[group.msgIds.length - 1],
+      chatUserId: group.userId,
+      replyCallback: async (replyContent: string) => {
+        await client.sendMessage({
+          to_user_id: group.userId,
+          text: replyContent,
+          context_token: group.contextToken
+        });
+        return { errcode: 0 };
+      }
+    });
+  } catch (error) {
+    // 错误处理（原有逻辑不变）
+    ...
+  }
+}
+```
+---
+## 六、文件大小与类型限制
+| 配置项 | 建议值 | 说明 |
+|--------|--------|------|
+| 单文件最大 | 20 MB | WeChat 个人号文件上限通常 ≤ 20MB |
+| 单消息最多文件 | 5 个 | 防止并发过多 |
+| 允许图片 MIME | `image/*` | 统一走图片存储 |
+| 允许文件类型 | 不限制（由 workflow 处理） | GridFS 存储，workflow 层可再过滤 |
+| 下载超时 | 30s | 防止 CDN 链接失效导致卡死 |
+---
+## 七、错误处理策略
+| 场景 | 处理方式 |
+|------|----------|
+| CDN URL 过期 / 无效 | 跳过该文件，继续处理文本；文件全部失败时回复提示 |
+| 文件过大 | 抛出错误，日志记录，回复"文件过大"提示 |
+| 图片 MIME 不合法 | `uploadMongoImg` 内部校验，返回 null，跳过 |
+| 单文件下载超时 | AbortController 控制，记录日志后 null |
+| 全部文件失败但有文本 | 仍走 workflow，文件部分缺失 |
+| 全部内容失败 | 发送固定错误消息给用户 |
+---
+## 八、待确认事项
+在开始编码前，需要确认以下 iLink API 细节：
+1. **图片消息 item type 值**：是否为 `2`？字段名是否为 `image_item`？
+2. **文件消息 item type 值**：是否为 `5`？字段名是否为 `file_item`？
+3. **CDN URL 有效期**：URL 可以直接 fetch 还是需要携带 token 鉴权？
+4. **是否有专用下载 API**：是否有类似 `/ilink/bot/getmedia?file_id=xxx` 的接口？
+5. **文件大小限制**：iLink 对单消息文件大小有何限制？
+---
+## 九、新增文件清单
+```
+packages/service/support/outLink/wechat/
+├── ilinkClient.ts          修改：扩展 MessageItem 类型，新增 downloadMedia()
+├── messageParser.ts        修改：新增 extractMediaRefFromItem()，扩展 ParsedMessageGroup
+├── fileHandler.ts          新建：下载 + 存储媒体文件
+├── mq.ts                   修改：调用 fileHandler，拼装 query
+└── type.ts                 保持不变
+```
+---
+## 十、测试用例设计
+| 用例 | 输入 | 期望输出 |
+|------|------|----------|
+| 纯文字消息 | text item | 不变，query=[text] |
+| 图片消息 | image item + 无文本 | query=[file(image)] |
+| 图片 + 文字 | text + image | query=[text, file(image)] |
+| 文件附件 | file item | query=[file(file)] |
+| 多文件 | 3 个图片 | query=[file, file, file] |
+| 超大文件 | 文件 > 20MB | 跳过，回复提示 |
+| CDN 失效 | 404 URL | 跳过，文本正常走 |
+| 全部文件失败 + 无文本 | 空 | 发送错误提示 |
--- a/document/content/docs/self-host/upgrading/4-14/41410.mdx
+++ b/document/content/docs/self-host/upgrading/4-14/41410.mdx
@@ -6,6 +6,7 @@ description: 'FastGPT V4.14.10 更新说明'
 ## 🚀 新增内容
+1. 飞书发布渠道，支持流输出。
 ## ⚙️ 优化

--- a/document/data/doc-last-modified.json
+++ b/document/data/doc-last-modified.json
@@ -220,6 +220,7 @@
  "document/content/docs/self-host/upgrading/4-14/4140.mdx": "2026-03-03T17:39:47+08:00",
  "document/content/docs/self-host/upgrading/4-14/4141.en.mdx": "2026-03-03T17:39:47+08:00",
  "document/content/docs/self-host/upgrading/4-14/4141.mdx": "2026-03-03T17:39:47+08:00",
+  "document/content/docs/self-host/upgrading/4-14/41410.mdx": "2026-03-24T09:59:30+08:00",
  "document/content/docs/self-host/upgrading/4-14/4142.en.mdx": "2026-03-03T17:39:47+08:00",
  "document/content/docs/self-host/upgrading/4-14/4142.mdx": "2026-03-03T17:39:47+08:00",
  "document/content/docs/self-host/upgrading/4-14/4143.en.mdx": "2026-03-03T17:39:47+08:00",
@@ -239,7 +240,7 @@
  "document/content/docs/self-host/upgrading/4-14/41481.en.mdx": "2026-03-09T12:02:02+08:00",
  "document/content/docs/self-host/upgrading/4-14/41481.mdx": "2026-03-09T17:39:53+08:00",
  "document/content/docs/self-host/upgrading/4-14/4149.en.mdx": "2026-03-23T12:17:04+08:00",
-  "document/content/docs/self-host/upgrading/4-14/4149.mdx": "2026-03-23T23:36:27+08:00",
+  "document/content/docs/self-host/upgrading/4-14/4149.mdx": "2026-03-23T23:37:12+08:00",
  "document/content/docs/self-host/upgrading/outdated/40.en.mdx": "2026-03-03T17:39:47+08:00",
  "document/content/docs/self-host/upgrading/outdated/40.mdx": "2026-03-03T17:39:47+08:00",
  "document/content/docs/self-host/upgrading/outdated/41.en.mdx": "2026-03-03T17:39:47+08:00",
@@ -381,7 +382,7 @@
  "document/content/docs/self-host/upgrading/upgrade-intruction.en.mdx": "2026-03-03T17:39:47+08:00",
  "document/content/docs/self-host/upgrading/upgrade-intruction.mdx": "2026-03-03T17:39:47+08:00",
  "document/content/docs/toc.en.mdx": "2026-03-23T12:10:44+08:00",
-  "document/content/docs/toc.mdx": "2026-03-23T11:57:05+08:00",
+  "document/content/docs/toc.mdx": "2026-03-24T09:59:30+08:00",
  "document/content/docs/use-cases/app-cases/dalle3.en.mdx": "2026-02-26T22:14:30+08:00",
  "document/content/docs/use-cases/app-cases/dalle3.mdx": "2025-07-23T21:35:03+08:00",
  "document/content/docs/use-cases/app-cases/english_essay_correction_bot.en.mdx": "2026-02-26T22:14:30+08:00",

--- a/packages/global/common/error/utils.ts
+++ b/packages/global/common/error/utils.ts
@@ -28,6 +28,10 @@ export const getErrText = (err: any, def = ''): any => {
  return replaceSensitiveText(msg);
 };
+export const getErrResponse = (err: any): any => {
+  return err?.response?.data || err?.response || err;
+};
 export class UserError extends Error {
  constructor(message: string) {
    super(message);

--- a/packages/service/common/api/axios.ts
+++ b/packages/service/common/api/axios.ts
 import _, { type AxiosRequestConfig } from 'axios';
 import { ProxyAgent } from 'proxy-agent';
+import { isDevEnv } from '@fastgpt/global/common/system/constants';
 export function createProxyAxios(config?: AxiosRequestConfig) {
+  if (isDevEnv) {
+    return _.create(config);
+  }
  const agent = new ProxyAgent();
  return _.create({

--- a/packages/service/support/outLink/runtime/utils.ts
+++ b/packages/service/support/outLink/runtime/utils.ts
@@ -21,7 +21,7 @@ import { authOutLinkLimit } from './auth';
 import { addOutLinkUsage } from '../../../support/outLink/tools';
 import { getLogger, LogCategories } from '../../../common/logger';
 import { appendRedisCache } from '../../../common/redis/cache';
-import { getErrText } from '@fastgpt/global/common/error/utils';
+import { getErrResponse, getErrText } from '@fastgpt/global/common/error/utils';
 import { getUsageSourceByPublishChannel } from '@fastgpt/global/support/wallet/usage/tools';
 import { getChatSourceByPublishChannel } from '@fastgpt/global/core/chat/utils';
 import { WORKFLOW_MAX_RUN_TIMES } from '../../../core/workflow/constants';
@@ -75,7 +75,10 @@ export type outLinkInvokeChatProps<T extends OutlinkAppType> = {
  res?: NextApiResponse;
  messageId: string;
  chatUserId: string;
-  replyCallback?: (replyContent: string) => Promise<any>;
+  // Called once with complete response content (all channels except wecom)
+  onReply?: (replyContent: string) => Promise<void>;
+  // Called for each streaming chunk (feishu and other push channels)
+  onStreamChunk?: (text: string) => Promise<void>;
  streamId?: string;
 };
@@ -91,7 +94,8 @@ export async function outlinkInvokeChat<T extends OutlinkAppType>({
  res,
  messageId,
  chatUserId,
-  replyCallback,
+  onReply,
+  onStreamChunk,
  streamId
 }: outLinkInvokeChatProps<T>) {
  const streamResKey = `${STREAM_CACHE_KEY_PREFIX}${streamId}`;
@@ -112,7 +116,7 @@ export async function outlinkInvokeChat<T extends OutlinkAppType>({
    const userQuestion = query.find((item) => item.text)?.text?.content || '';
    if (RESET_CHAT_INPUT[userQuestion]) {
      await resetChat({ appId: outLinkConfig.appId, chatId });
-      await replyCallback?.(RESET_CHAT_REPLY);
+      await onReply?.(RESET_CHAT_REPLY);
      if (streamId) {
        await appendRedisCache(streamResKey, RESET_CHAT_REPLY, 60);
        await appendRedisCache(streamResKey, STREAM_END_FLAG, 60);
@@ -144,7 +148,9 @@ export async function outlinkInvokeChat<T extends OutlinkAppType>({
      ip: chatId
    });
-    const workflowStreamResponse = streamId
+    const enableStreaming = !!streamId || !!onStreamChunk;
+    const workflowStreamResponse = enableStreaming
      ? async ({
          write,
          event,
@@ -158,8 +164,13 @@ export async function outlinkInvokeChat<T extends OutlinkAppType>({
            try {
              const text = data.choices?.[0]?.delta?.content;
              if (text) {
+                if (streamId) {
                  await appendRedisCache(streamResKey, text, 60);
                }
+                if (onStreamChunk) {
+                  await onStreamChunk(text);
+                }
+              }
            } catch (error) {
              logger.error('Outlink real-time streaming failed', {
                streamId,
@@ -171,8 +182,10 @@ export async function outlinkInvokeChat<T extends OutlinkAppType>({
        }
      : undefined;
-    // Append first
+    // Initialize Redis key only when needed
+    if (streamId) {
      await appendRedisCache(streamResKey, '', 120);
+    }
    // Merge global variables from database
    const variables = chatDetail?.variables ?? {};
@@ -202,7 +215,7 @@ export async function outlinkInvokeChat<T extends OutlinkAppType>({
      histories,
      query: query,
      chatConfig,
-      stream: streamId ? true : false,
+      stream: enableStreaming,
      workflowStreamResponse,
      runtimeEdges: storeEdges2RuntimeEdges(edges),
      runtimeNodes: storeNodes2RuntimeNodes(nodes, getWorkflowEntryNodeIds(nodes)),
@@ -225,31 +238,13 @@ export async function outlinkInvokeChat<T extends OutlinkAppType>({
    const replyResult = await (async () => {
      try {
        if (streamId) {
-          // if streamId is provided, do not reply
+          // wecom model: Redis handles streaming, no reply needed
-          return;
+          return { success: true };
-        }
-        const result = await replyCallback?.(responseContent);
-        if (result.errcode !== 0) {
-          logger.error('Outlink official account reply failed', {
-            errmsg: result.errmsg
-          });
-          return {
-            success: false,
-            errmsg: result.errmsg
-          };
        }
+        await onReply?.(responseContent);
-        logger.debug('Outlink official account reply success', {
+        return { success: true };
-          responseContent,
-          result
-        });
-        return {
-          success: true,
-          data: result
-        };
      } catch (error) {
-        logger.error('Outlink reply callback failed', { error });
+        logger.error('Outlink reply callback failed', { error: getErrResponse(error) });
        return {
          success: false,
          errmsg: getErrText(error)
@@ -306,9 +301,10 @@ export async function outlinkInvokeChat<T extends OutlinkAppType>({
    });
    try {
+      if (streamId) {
        await appendRedisCache(streamResKey, STREAM_END_FLAG, 60);
+      }
-      await replyCallback?.(`App run error: ${getErrText(error)}`);
+      await onReply?.(`App run error: ${getErrText(error)}`);
    } catch (error) {
      logger.error('Outlink invoke chat fallback reply failed', {
        shareId: outLinkConfig.shareId,

--- a/packages/service/support/outLink/wechat/mq.ts
+++ b/packages/service/support/outLink/wechat/mq.ts
@@ -126,13 +126,12 @@ async function processUserGroup(
      query: [{ text: { content: group.text } }],
      messageId: group.msgIds[group.msgIds.length - 1],
      chatUserId: group.userId,
-      replyCallback: async (replyContent: string) => {
+      onReply: async (replyContent: string) => {
        await client.sendMessage({
          to_user_id: group.userId,
          text: replyContent,
          context_token: group.contextToken
        });
-        return { errcode: 0 };
      }
    });
  } catch (error) {