✅ test: add e2e regression test for topic switch page reload

Add a regression test to verify that switching topics within the same agent does not trigger a full browser page reload. The test injects a window marker before switching and verifies it survives after the switch. Related: lobehub/lobehub#13309 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
🐛 fix: prevent full page reload when switching topics during agent execution (#13309 )
2026-06-15 12:10:16 +00:00 · 2026-03-26 21:31:19 +08:00 · 2026-03-26 21:30:08 +08:00 · 2026-03-26 21:16:38 +08:00 · 2026-03-26 21:13:55 +08:00 · 2026-03-26 20:41:06 +08:00
100 changed files with 2042 additions and 1160 deletions
@@ -0,0 +1,15 @@
+@regression @P0 @agent @topic-switch
+Feature: 切换话题不触发页面全量刷新
+  作为用户，当我在 Agent 执行过程中切换话题时，页面不应该发生全量重新加载
+
+  Background:
+    Given 用户已登录系统
+    And 用户进入 Lobe AI 对话页面
+
+  @AGENT-TOPIC-RELOAD-001 @smoke
+  Scenario: 在同一 Agent 下切换话题不触发页面刷新
+    Given 用户在当前 Agent 中创建了两个对话
+    When 用户在页面注入状态标记
+    And 用户切换到另一个话题
+    Then 页面状态标记应该仍然存在
+    And 页面导航类型不应该是全量加载
@@ -0,0 +1,168 @@
+/**
+ * Topic Switch No-Reload Regression Test Steps
+ *
+ * Verifies that switching topics within the same agent does NOT trigger a full
+ * browser page reload. The bug was caused by NavItem's onClick handler skipping
+ * e.preventDefault() when disabled/loading was true, allowing the <a> tag's
+ * default navigation to fire.
+ *
+ * Detection: inject a window marker before switching, verify it survives.
+ */
+import { Given, Then, When } from '@cucumber/cucumber';
+import { expect } from '@playwright/test';
+
+import type { CustomWorld } from '../../support/world';
+import { WAIT_TIMEOUT } from '../../support/world';
+
+// ============================================
+// Helpers
+// ============================================
+
+async function focusAndType(world: CustomWorld, text: string): Promise<void> {
+  const candidates = [
+    world.page.locator(
+      'textarea[placeholder*="Ask"], textarea[placeholder*="Press"], textarea[placeholder*="输入"]',
+    ),
+    world.page.locator('[data-testid="chat-input"] [contenteditable="true"]'),
+    world.page.locator('[data-testid="chat-input"] textarea'),
+  ];
+
+  for (const locator of candidates) {
+    const count = await locator.count();
+    for (let i = 0; i < count; i++) {
+      const item = locator.nth(i);
+      if (await item.isVisible().catch(() => false)) {
+        await item.click({ force: true });
+        await world.page.waitForTimeout(300);
+        await world.page.keyboard.type(text, { delay: 30 });
+        return;
+      }
+    }
+  }
+
+  throw new Error('Could not find a visible chat input');
+}
+
+// ============================================
+// Given Steps
+// ============================================
+
+Given('用户在当前 Agent 中创建了两个对话', async function (this: CustomWorld) {
+  console.log('   📍 Step: 创建第一个对话...');
+
+  // Send first message to create topic 1
+  await focusAndType(this, 'hello');
+  await this.page.keyboard.press('Enter');
+  await this.page.waitForTimeout(3000);
+
+  // Verify a topic appeared in sidebar
+  const topicItems = this.page.locator('[data-testid="topic-item"]');
+  await expect(topicItems.first()).toBeVisible({ timeout: WAIT_TIMEOUT });
+  console.log('   ✅ 第一个对话已创建');
+
+  // Create a new topic
+  console.log('   📍 Step: 创建第二个对话...');
+  const addTopicButton = this.page.locator('svg.lucide-message-square-plus').locator('..');
+  await expect(addTopicButton.first()).toBeVisible({ timeout: 5000 });
+  await addTopicButton.first().click();
+  await this.page.waitForTimeout(1000);
+
+  // Send message to create topic 2
+  await focusAndType(this, 'world');
+  await this.page.keyboard.press('Enter');
+  await this.page.waitForTimeout(3000);
+
+  // Verify we now have at least 2 topics
+  const topicCount = await topicItems.count();
+  console.log(`   📍 话题数量: ${topicCount}`);
+  expect(topicCount).toBeGreaterThanOrEqual(2);
+  console.log('   ✅ 两个对话已创建');
+});
+
+// ============================================
+// When Steps
+// ============================================
+
+When('用户在页面注入状态标记', async function (this: CustomWorld) {
+  console.log('   📍 Step: 注入页面状态标记...');
+
+  await this.page.evaluate(() => {
+    (window as any).__e2eNoReloadMarker = true;
+  });
+
+  // Verify marker was set
+  const marker = await this.page.evaluate(() => (window as any).__e2eNoReloadMarker);
+  expect(marker).toBe(true);
+  console.log('   ✅ 状态标记已注入');
+});
+
+When('用户切换到另一个话题', async function (this: CustomWorld) {
+  console.log('   📍 Step: 切换到另一个话题...');
+
+  const topicItems = this.page.locator('[data-testid="topic-item"]');
+  const topicCount = await topicItems.count();
+  console.log(`   📍 找到 ${topicCount} 个话题`);
+
+  // Find the first non-active topic and click it
+  for (let i = 0; i < topicCount; i++) {
+    const topic = topicItems.nth(i);
+    // Check if this topic is NOT currently active (doesn't have active/filled variant)
+    const isActive = await topic.evaluate((el) => {
+      // Walk up to find the NavItem wrapper and check its variant/active state
+      const navItem = el.closest('[class*="Block"]');
+      return navItem?.getAttribute('data-active') === 'true' || el.classList.contains('active');
+    });
+
+    if (!isActive) {
+      await topic.click();
+      console.log(`   ✅ 已点击第 ${i + 1} 个话题`);
+      await this.page.waitForTimeout(2000);
+      return;
+    }
+  }
+
+  // Fallback: just click the first topic
+  await topicItems.first().click();
+  console.log('   ✅ 已点击第一个话题（fallback）');
+  await this.page.waitForTimeout(2000);
+});
+
+// ============================================
+// Then Steps
+// ============================================
+
+Then('页面状态标记应该仍然存在', async function (this: CustomWorld) {
+  console.log('   📍 Step: 验证页面状态标记...');
+
+  const marker = await this.page.evaluate(() => (window as any).__e2eNoReloadMarker);
+
+  if (marker !== true) {
+    // Take screenshot for debugging
+    await this.takeScreenshot('topic-switch-reload-detected');
+    throw new Error(
+      'Page reload detected! window.__e2eNoReloadMarker was lost after topic switch. ' +
+        'This means the <a> tag default navigation fired instead of SPA routing.',
+    );
+  }
+
+  console.log('   ✅ 页面状态标记仍然存在（未发生 reload）');
+});
+
+Then('页面导航类型不应该是全量加载', async function (this: CustomWorld) {
+  console.log('   📍 Step: 检查导航类型...');
+
+  const navInfo = await this.page.evaluate(() => {
+    const entry = performance.getEntriesByType('navigation')[0] as PerformanceNavigationTiming;
+    return {
+      duration: entry?.duration,
+      type: entry?.type,
+    };
+  });
+
+  console.log(`   📍 Navigation type: ${navInfo.type}, duration: ${navInfo.duration}ms`);
+
+  // If a full page reload happened AFTER the initial load, the navigation type
+  // would still show the initial load type. The window marker check above is
+  // the primary detection. This step provides additional diagnostic info.
+  console.log('   ✅ 导航信息已记录');
+});
@@ -4,6 +4,9 @@
  "error.retry": "Reload",
  "error.stack": "Error Stack",
  "error.title": "Oops, something went wrong..",
+  "exceededContext.compact": "Compact Context",
+  "exceededContext.desc": "The conversation has exceeded the context window limit. You can compact the context to compress history and continue chatting.",
+  "exceededContext.title": "Context Window Exceeded",
  "fetchError.detail": "Error details",
  "fetchError.title": "Request failed",
  "import.importConfigFile.description": "Error reason: {{reason}}",
@@ -108,7 +111,7 @@
  "response.PluginSettingsInvalid": "This skill needs to be correctly configured before it can be used. Please check if your configuration is correct",
  "response.ProviderBizError": "Error requesting {{provider}} service, please troubleshoot or retry based on the following information",
  "response.QuotaLimitReached": "Sorry, the token usage or request count has reached the quota limit for this key. Please increase the key's quota or try again later.",
-  "response.QuotaLimitReachedCloud": "The model service is currently under heavy load. Please try again later.",
+  "response.QuotaLimitReachedCloud": "The model service is currently under heavy load. Please try again later or switch to another model.",
  "response.ServerAgentRuntimeError": "Sorry, the Agent service is currently unavailable. Please try again later or contact us via email for support.",
  "response.StreamChunkError": "Error parsing the message chunk of the streaming request. Please check if the current API interface complies with the standard specifications, or contact your API provider for assistance.",
  "response.SubscriptionKeyMismatch": "We apologize for the inconvenience. Due to a temporary system malfunction, your current subscription usage is inactive. Please click the button below to restore your subscription, or contact us via email for support.",
@@ -120,6 +123,10 @@
  "supervisor.decisionFailed": "The group host is unable to function. Please check your host configuration to ensure the correct model, API Key, and API endpoint are set.",
  "testConnectionFailed": "Test connection failed: {{error}}",
  "tts.responseError": "Service request failed, please check the configuration or try again",
+  "unknownError.copyTraceId": "Trace ID Copied",
+  "unknownError.desc": "An unexpected error occurred. You can retry or report on",
+  "unknownError.retry": "Retry",
+  "unknownError.title": "Oops, the request took a nap",
  "unlock.addProxyUrl": "Add OpenAI proxy URL (optional)",
  "unlock.apiKey.description": "Enter your {{name}} API Key to start the session",
  "unlock.apiKey.imageGenerationDescription": "Enter your {{name}} API Key to start generating",
@@ -0,0 +1,12 @@
+{
+  "image_generation_completed": "Image \"{{prompt}}\" generated successfully",
+  "image_generation_completed_title": "Image Generated",
+  "inbox.archiveAll": "Archive all",
+  "inbox.empty": "No notifications yet",
+  "inbox.emptyUnread": "No unread notifications",
+  "inbox.filterUnread": "Show unread only",
+  "inbox.markAllRead": "Mark all as read",
+  "inbox.title": "Notifications",
+  "video_generation_completed": "Video \"{{prompt}}\" generated successfully",
+  "video_generation_completed_title": "Video Generated"
+}
@@ -443,6 +443,12 @@
  "myAgents.status.published": "Published",
  "myAgents.status.unpublished": "Unpublished",
  "myAgents.title": "My Published Agents",
+  "notification.email.desc": "Receive email notifications when important events occur",
+  "notification.email.title": "Email Notifications",
+  "notification.enabled": "Enabled",
+  "notification.inbox.desc": "Show notifications in the in-app inbox",
+  "notification.inbox.title": "Inbox Notifications",
+  "notification.title": "Notification Channels",
  "plugin.addMCPPlugin": "Add MCP",
  "plugin.addTooltip": "Custom Skills",
  "plugin.clearDeprecated": "Remove Deprecated Skills",
@@ -807,6 +813,7 @@
  "tab.manualFill": "Manually Fill In",
  "tab.manualFill.desc": "Configure a custom MCP skill manually",
  "tab.memory": "Memory",
+  "tab.notification": "Notifications",
  "tab.profile": "My Account",
  "tab.provider": "Provider",
  "tab.proxy": "Proxy",
@@ -4,6 +4,9 @@
  "error.retry": "重新加载",
  "error.stack": "错误堆栈",
  "error.title": "页面暂时不可用",
+  "exceededContext.compact": "压缩上下文",
+  "exceededContext.desc": "对话已超出模型上下文窗口限制。你可以压缩上下文来压缩历史记录并继续对话。",
+  "exceededContext.title": "上下文窗口超出限制",
  "fetchError.detail": "查看详情",
  "fetchError.title": "请求未能完成",
  "import.importConfigFile.description": "原因：{{reason}}",
@@ -108,7 +111,7 @@
  "response.PluginSettingsInvalid": "该技能需要完成配置后才能使用，请检查技能配置",
  "response.ProviderBizError": "模型服务商返回错误。请根据以下信息排查，或稍后重试",
  "response.QuotaLimitReached": "Token 用量或请求次数已达配额上限。请提升配额或稍后再试",
-  "response.QuotaLimitReachedCloud": "当前模型服务负载较高，请稍后重试。",
+  "response.QuotaLimitReachedCloud": "当前模型服务负载较高，请稍后重试或切换其他模型。",
  "response.ServerAgentRuntimeError": "助理运行服务暂不可用。请稍后再试，或邮件联系我们",
  "response.StreamChunkError": "流式响应解析失败。请检查接口是否符合规范，或联系模型服务商",
  "response.SubscriptionKeyMismatch": "订阅状态同步异常。你可以点击下方按钮恢复订阅，或邮件联系我们",
@@ -120,6 +123,10 @@
  "supervisor.decisionFailed": "群组主持人运行失败。请检查主持人配置（模型、API Key 与 API 地址）后重试",
  "testConnectionFailed": "测试连接失败：{{error}}",
  "tts.responseError": "请求失败。请检查配置后重试",
+  "unknownError.copyTraceId": "Trace ID 已复制",
+  "unknownError.desc": "遇到了意外错误，请重试或反馈至",
+  "unknownError.retry": "重试",
+  "unknownError.title": "糟糕，请求打了个盹",
  "unlock.addProxyUrl": "添加 OpenAI 代理地址（可选）",
  "unlock.apiKey.description": "输入你的 {{name}} API Key，即可开始会话",
  "unlock.apiKey.imageGenerationDescription": "输入你的 {{name}} API Key，即可开始生成",
@@ -0,0 +1,12 @@
+{
+  "image_generation_completed": "图片 \"{{prompt}}\" 已生成完成",
+  "image_generation_completed_title": "图片已生成",
+  "inbox.archiveAll": "全部归档",
+  "inbox.empty": "暂无通知",
+  "inbox.emptyUnread": "没有未读通知",
+  "inbox.filterUnread": "仅显示未读",
+  "inbox.markAllRead": "全部标为已读",
+  "inbox.title": "通知",
+  "video_generation_completed": "视频 \"{{prompt}}\" 已生成完成",
+  "video_generation_completed_title": "视频已生成"
+}
@@ -443,6 +443,12 @@
  "myAgents.status.published": "已上架",
  "myAgents.status.unpublished": "未上架",
  "myAgents.title": "我发布的助理",
+  "notification.email.desc": "当重要事件发生时接收邮件通知",
+  "notification.email.title": "邮件通知",
+  "notification.enabled": "启用",
+  "notification.inbox.desc": "在应用内收件箱中显示通知",
+  "notification.inbox.title": "站内通知",
+  "notification.title": "通知渠道",
  "plugin.addMCPPlugin": "添加 MCP",
  "plugin.addTooltip": "自定义技能",
  "plugin.clearDeprecated": "移除无效技能",
@@ -807,6 +813,7 @@
  "tab.manualFill": "自行填写内容",
  "tab.manualFill.desc": "手动配置自定义 MCP 技能",
  "tab.memory": "记忆设置",
+  "tab.notification": "通知",
  "tab.profile": "我的账号",
  "tab.provider": "AI 服务商",
  "tab.proxy": "网络代理",
@@ -196,7 +196,6 @@
    "@huggingface/inference": "^4.13.10",
    "@icons-pack/react-simple-icons": "^13.8.0",
    "@khmyznikov/pwa-install": "0.3.9",
-    "@langchain/community": "^0.3.59",
    "@lexical/utils": "^0.39.0",
    "@lobechat/agent-runtime": "workspace:*",
    "@lobechat/builtin-agents": "workspace:*",
@@ -308,6 +307,7 @@
    "cmdk": "^1.1.1",
    "cookie": "^1.1.1",
    "countries-and-timezones": "^3.8.0",
+    "d3-dsv": "^3.0.1",
    "dayjs": "^1.11.19",
    "debug": "^4.4.3",
    "dexie": "^3.2.7",
@@ -333,7 +333,6 @@
    "js-sha256": "^0.11.1",
    "jsonl-parse-stringify": "^1.0.3",
    "klavis": "^2.15.0",
-    "langchain": "^0.3.37",
    "langfuse": "^3.38.6",
    "langfuse-core": "^3.38.6",
    "lexical": "^0.39.0",
@@ -402,7 +401,6 @@
    "superjson": "^2.2.6",
    "svix": "^1.84.1",
    "swr": "^2.3.8",
-    "systemjs": "^6.15.1",
    "three": "^0.181.2",
    "tokenx": "^1.3.0",
    "ts-md5": "^2.0.1",
@@ -444,21 +442,23 @@
    "@types/async-retry": "^1.4.9",
    "@types/chroma-js": "^3.1.2",
    "@types/crypto-js": "^4.2.2",
+    "@types/d3-dsv": "^3.0.7",
    "@types/debug": "^4.1.12",
    "@types/fs-extra": "^11.0.4",
+    "@types/html-to-text": "^9.0.4",
    "@types/ip": "^1.1.3",
    "@types/json-schema": "^7.0.15",
    "@types/node": "^24.10.9",
    "@types/nodemailer": "^7.0.5",
    "@types/numeral": "^2.0.5",
    "@types/oidc-provider": "^9.5.0",
+    "@types/pdf-parse": "^1.1.4",
    "@types/pdfkit": "^0.17.4",
    "@types/pg": "^8.16.0",
    "@types/react": "19.2.13",
    "@types/react-dom": "^19.2.3",
    "@types/rtl-detect": "^1.0.3",
    "@types/semver": "^7.7.1",
-    "@types/systemjs": "^6.15.4",
    "@types/three": "^0.181.0",
    "@types/ua-parser-js": "^0.7.39",
    "@types/unist": "^3.0.3",
@@ -5,6 +5,7 @@ import {
  DEFAULT_HOTKEY_CONFIG,
  DEFAULT_IMAGE_CONFIG,
  DEFAULT_MEMORY_SETTINGS,
+  DEFAULT_NOTIFICATION_SETTINGS,
  DEFAULT_SYSTEM_AGENT_CONFIG,
  DEFAULT_TOOL_CONFIG,
  DEFAULT_TTS_CONFIG,
@@ -19,6 +20,7 @@ export const DEFAULT_SETTINGS: UserSettings = {
  keyVaults: {},
  languageModel: DEFAULT_LLM_CONFIG,
  memory: DEFAULT_MEMORY_SETTINGS,
+  notification: DEFAULT_NOTIFICATION_SETTINGS,
  systemAgent: DEFAULT_SYSTEM_AGENT_CONFIG,
  tool: DEFAULT_TOOL_CONFIG,
  tts: DEFAULT_TTS_CONFIG,
@@ -6,6 +6,7 @@ export * from './image';
 export * from './knowledge';
 export * from './llm';
 export * from './memory';
+export * from './notification';
 export * from './systemAgent';
 export * from './tool';
 export * from './tts';
@@ -0,0 +1,22 @@
+import type { NotificationSettings } from '@lobechat/types';
+
+export const DEFAULT_NOTIFICATION_SETTINGS: NotificationSettings = {
+  email: {
+    enabled: true,
+    items: {
+      generation: {
+        image_generation_completed: true,
+        video_generation_completed: true,
+      },
+    },
+  },
+  inbox: {
+    enabled: true,
+    items: {
+      generation: {
+        image_generation_completed: true,
+        video_generation_completed: true,
+      },
+    },
+  },
+};
@@ -0,0 +1,128 @@
+import { and, count, desc, eq, inArray, lt, or } from 'drizzle-orm';
+
+import type { NewNotification, NewNotificationDelivery } from '../schemas/notification';
+import { notificationDeliveries, notifications } from '../schemas/notification';
+import type { LobeChatDatabase } from '../type';
+
+export class NotificationModel {
+  private readonly userId: string;
+  private readonly db: LobeChatDatabase;
+
+  constructor(db: LobeChatDatabase, userId: string) {
+    this.db = db;
+    this.userId = userId;
+  }
+
+  async list(
+    opts: { category?: string; cursor?: string; limit?: number; unreadOnly?: boolean } = {},
+  ) {
+    const { cursor, limit = 20, category, unreadOnly } = opts;
+
+    const conditions = [eq(notifications.userId, this.userId), eq(notifications.isArchived, false)];
+
+    if (unreadOnly) {
+      conditions.push(eq(notifications.isRead, false));
+    }
+
+    if (category) {
+      conditions.push(eq(notifications.category, category));
+    }
+
+    if (cursor) {
+      const cursorRow = await this.db
+        .select({ createdAt: notifications.createdAt, id: notifications.id })
+        .from(notifications)
+        .where(and(eq(notifications.id, cursor), eq(notifications.userId, this.userId)))
+        .limit(1);
+
+      if (cursorRow[0]) {
+        // Composite cursor to handle identical createdAt timestamps
+        const { createdAt: cursorTime, id: cursorId } = cursorRow[0];
+        conditions.push(
+          or(
+            lt(notifications.createdAt, cursorTime),
+            and(eq(notifications.createdAt, cursorTime), lt(notifications.id, cursorId)),
+          )!,
+        );
+      }
+    }
+
+    return this.db
+      .select()
+      .from(notifications)
+      .where(and(...conditions))
+      .orderBy(desc(notifications.createdAt), desc(notifications.id))
+      .limit(limit);
+  }
+
+  async getUnreadCount(): Promise<number> {
+    const [result] = await this.db
+      .select({ count: count() })
+      .from(notifications)
+      .where(
+        and(
+          eq(notifications.userId, this.userId),
+          eq(notifications.isRead, false),
+          eq(notifications.isArchived, false),
+        ),
+      );
+
+    return result?.count ?? 0;
+  }
+
+  async markAsRead(ids: string[]) {
+    if (ids.length === 0) return;
+
+    return this.db
+      .update(notifications)
+      .set({ isRead: true, updatedAt: new Date() })
+      .where(and(eq(notifications.userId, this.userId), inArray(notifications.id, ids)));
+  }
+
+  async markAllAsRead() {
+    return this.db
+      .update(notifications)
+      .set({ isRead: true, updatedAt: new Date() })
+      .where(
+        and(
+          eq(notifications.userId, this.userId),
+          eq(notifications.isRead, false),
+          eq(notifications.isArchived, false),
+        ),
+      );
+  }
+
+  async archive(id: string) {
+    return this.db
+      .update(notifications)
+      .set({ isArchived: true, updatedAt: new Date() })
+      .where(and(eq(notifications.id, id), eq(notifications.userId, this.userId)));
+  }
+
+  async archiveAll() {
+    return this.db
+      .update(notifications)
+      .set({ isArchived: true, updatedAt: new Date() })
+      .where(and(eq(notifications.userId, this.userId), eq(notifications.isArchived, false)));
+  }
+
+  // ─── Write-side (used by NotificationService in cloud) ─────────
+
+  async create(data: Omit<NewNotification, 'userId'>) {
+    const [result] = await this.db
+      .insert(notifications)
+      .values({ ...data, userId: this.userId })
+      .onConflictDoNothing({
+        target: [notifications.userId, notifications.dedupeKey],
+      })
+      .returning();
+
+    return result ?? null;
+  }
+
+  async createDelivery(data: NewNotificationDelivery) {
+    const [result] = await this.db.insert(notificationDeliveries).values(data).returning();
+
+    return result;
+  }
+}
@@ -1,3 +1,4 @@
+import { AgentRuntimeErrorType } from '@lobechat/types';
 import { beforeEach, describe, expect, it, vi } from 'vitest';

 import type { LobeRuntimeAI } from '../BaseAI';
@@ -421,6 +422,83 @@ describe('createRouterRuntime', () => {
      ).rejects.toThrow('empty provider options');
    });

+    it('should not retry when ExceededContextWindow error is thrown', async () => {
+      const exceededError = {
+        errorType: AgentRuntimeErrorType.ExceededContextWindow,
+        error: { message: 'Too many input tokens' },
+        provider: 'test',
+      };
+
+      const mockChatFail = vi.fn().mockRejectedValue(exceededError);
+      const mockChatSuccess = vi.fn().mockResolvedValue('success');
+
+      class FailRuntime implements LobeRuntimeAI {
+        chat = mockChatFail;
+      }
+
+      class SuccessRuntime implements LobeRuntimeAI {
+        chat = mockChatSuccess;
+      }
+
+      const Runtime = createRouterRuntime({
+        id: 'test-runtime',
+        routers: [
+          {
+            apiType: 'openai',
+            options: [
+              { apiKey: 'key-1', runtime: FailRuntime as any },
+              { apiKey: 'key-2', runtime: SuccessRuntime as any },
+            ],
+            runtime: FailRuntime as any,
+            models: ['gpt-4'],
+          },
+        ],
+      });
+
+      const runtime = new Runtime();
+      await expect(
+        runtime.chat({ model: 'gpt-4', messages: [], temperature: 0.7 }),
+      ).rejects.toEqual(exceededError);
+
+      // Second channel should never be called
+      expect(mockChatFail).toHaveBeenCalledTimes(1);
+      expect(mockChatSuccess).not.toHaveBeenCalled();
+    });
+
+    it('should still retry on other error types', async () => {
+      const bizError = {
+        errorType: AgentRuntimeErrorType.ProviderBizError,
+        error: { message: 'Server error' },
+        provider: 'test',
+      };
+
+      const mockChatFail = vi.fn().mockRejectedValue(bizError);
+
+      class FailRuntime implements LobeRuntimeAI {
+        chat = mockChatFail;
+      }
+
+      const Runtime = createRouterRuntime({
+        id: 'test-runtime',
+        routers: [
+          {
+            apiType: 'openai',
+            options: [{ apiKey: 'key-1' }, { apiKey: 'key-2' }],
+            runtime: FailRuntime as any,
+            models: ['gpt-4'],
+          },
+        ],
+      });
+
+      const runtime = new Runtime();
+      await expect(
+        runtime.chat({ model: 'gpt-4', messages: [], temperature: 0.7 }),
+      ).rejects.toEqual(bizError);
+
+      // Both channels should be tried
+      expect(mockChatFail).toHaveBeenCalledTimes(2);
+    });
+
    it('should use apiType from option item when specified for fallback', async () => {
      const constructorCalls: any[] = [];

@@ -2,7 +2,7 @@
 * @see https://github.com/lobehub/lobe-chat/discussions/6563
 */
 import type { GoogleGenAIOptions } from '@google/genai';
-import type { ChatModelCard } from '@lobechat/types';
+import { AgentRuntimeErrorType, type ChatModelCard } from '@lobechat/types';
 import debug from 'debug';
 import type { ClientOptions } from 'openai';
 import type OpenAI from 'openai';
@@ -392,6 +392,14 @@ export const createRouterRuntime = ({
              log('onRouteAttempt callback error: %O', e);
            });

+          // Non-retryable errors: the request itself is invalid, retrying with another channel won't help
+          if (
+            (error as ChatCompletionErrorPayload)?.errorType ===
+            AgentRuntimeErrorType.ExceededContextWindow
+          ) {
+            throw error;
+          }
+
          if (attempt < totalOptions) {
            log(
              'attempt %d/%d failed (model=%s apiType=%s channelId=%s remark=%s), trying next',
@@ -703,6 +703,29 @@ describe('LobeBedrockAI', () => {
          }),
        );
      });
+
+      it('should throw ExceededContextWindow when error message indicates context window exceeded', async () => {
+        const errorMessage =
+          'Too many input tokens. Max input tokens for this model is 200000, but 250000 were provided.';
+        const errorMetadata = { statusCode: 400 };
+        const mockError = new Error(errorMessage);
+        (mockError as any).$metadata = errorMetadata;
+        (instance['client'].send as Mock).mockRejectedValue(mockError);
+
+        await expect(
+          instance.chat({
+            max_tokens: 100,
+            messages: [{ content: 'Hello', role: 'user' }],
+            model: 'anthropic.claude-v2:1',
+            temperature: 0,
+          }),
+        ).rejects.toThrow(
+          expect.objectContaining({
+            errorType: AgentRuntimeErrorType.ExceededContextWindow,
+            provider: ModelProvider.Bedrock,
+          }),
+        );
+      });
    });

    describe('Llama Model', () => {
@@ -28,6 +28,7 @@ import { AgentRuntimeErrorType } from '../../types/error';
 import { AgentRuntimeError } from '../../utils/createError';
 import { debugStream } from '../../utils/debugStream';
 import { getModelPricing } from '../../utils/getModelPricing';
+import { isExceededContextWindowError } from '../../utils/isExceededContextWindowError';
 import { StreamingResponse } from '../../utils/response';

 /**
@@ -284,6 +285,9 @@ export class LobeBedrockAI implements LobeRuntimeAI {
      );
    } catch (e) {
      const err = e as Error & { $metadata: any };
+      const errorType = isExceededContextWindowError(err.message)
+        ? AgentRuntimeErrorType.ExceededContextWindow
+        : AgentRuntimeErrorType.ProviderBizError;

      throw AgentRuntimeError.chat({
        error: {
@@ -291,7 +295,7 @@ export class LobeBedrockAI implements LobeRuntimeAI {
          message: err.message,
          type: err.name,
        },
-        errorType: AgentRuntimeErrorType.ProviderBizError,
+        errorType,
        provider: this.id,
        region: this.region,
      });
@@ -103,6 +103,22 @@ export class LobeFalAI implements LobeRuntimeAI {
        });
      }

+      // 422 ValidationError with content_policy_violation — show a clean message
+      if (error instanceof Error && 'status' in error && error.status === 422) {
+        const body = 'body' in error ? (error as any).body : undefined;
+        const hasContentPolicyViolation =
+          Array.isArray(body?.detail) &&
+          body.detail.some((d: any) => d.type === 'content_policy_violation');
+
+        if (hasContentPolicyViolation) {
+          throw AgentRuntimeError.createError(AgentRuntimeErrorType.ProviderBizError, {
+            error,
+            message:
+              'The request content violates content policy. Please modify your prompt and try again.',
+          });
+        }
+      }
+
      throw AgentRuntimeError.createError(AgentRuntimeErrorType.ProviderBizError, { error });
    }
  }
@@ -8,6 +8,7 @@ import type { UserKeyVaults } from './keyVaults';
 import type { MarketAuthTokens } from './market';
 import type { UserMemorySettings } from './memory';
 import type { UserModelProviderConfig } from './modelProvider';
+import type { NotificationSettings } from './notification';
 import type { UserSystemAgentConfig } from './systemAgent';
 import type { UserToolConfig } from './tool';
 import type { UserTTSConfig } from './tts';
@@ -22,6 +23,7 @@ export * from './keyVaults';
 export * from './market';
 export * from './memory';
 export * from './modelProvider';
+export * from './notification';
 export * from './sync';
 export * from './systemAgent';
 export * from './tool';
@@ -39,6 +41,7 @@ export interface UserSettings {
  languageModel: UserModelProviderConfig;
  market?: MarketAuthTokens;
  memory?: UserMemorySettings;
+  notification?: NotificationSettings;
  systemAgent: UserSystemAgentConfig;
  tool: UserToolConfig;
  tts: UserTTSConfig;
@@ -58,6 +61,7 @@ export const UserSettingsSchema = z
    languageModel: z.any().optional(),
    market: z.any().optional(),
    memory: z.any().optional(),
+    notification: z.any().optional(),
    systemAgent: z.any().optional(),
    tool: z.any().optional(),
    tts: z.any().optional(),
@@ -0,0 +1,10 @@
+export interface NotificationChannelSettings {
+  enabled?: boolean;
+  /** Per-type overrides grouped by category. Missing = use scenario default (true) */
+  items?: Record<string, Record<string, boolean>>;
+}
+
+export interface NotificationSettings {
+  email?: NotificationChannelSettings;
+  inbox?: NotificationChannelSettings;
+}
@@ -15,6 +15,7 @@ import { type RuntimeVideoGenParams } from 'model-bank';
 import { NextResponse } from 'next/server';

 import { chargeAfterGenerate } from '@/business/server/video-generation/chargeAfterGenerate';
+import { notifyVideoCompleted } from '@/business/server/video-generation/notifyVideoCompleted';
 import { AsyncTaskModel } from '@/database/models/asyncTask';
 import { GenerationModel } from '@/database/models/generation';
 import { generationBatches } from '@/database/schemas';
@@ -201,6 +202,14 @@ export const POST = async (req: Request, { params }: { params: Promise<{ provide
      status: AsyncTaskStatus.Success,
    });

+    notifyVideoCompleted({
+      generationBatchId: generation.generationBatchId!,
+      model: resolvedModel,
+      prompt: batch?.prompt ?? '',
+      topicId: batch?.generationTopicId,
+      userId: asyncTask.userId,
+    }).catch((err) => console.error('[video-webhook] notification failed:', err));
+
    // Charge after successful video generation
    try {
      await chargeAfterGenerate({
@@ -0,0 +1,3 @@
+const Notification = () => null;
+
+export default Notification;
@@ -0,0 +1,11 @@
+interface NotifyImageCompletedParams {
+  duration: number;
+  generationBatchId: string;
+  model: string;
+  prompt: string;
+  topicId?: string;
+  userId: string;
+}
+
+// eslint-disable-next-line unused-imports/no-unused-vars
+export async function notifyImageCompleted(params: NotifyImageCompletedParams): Promise<void> {}
@@ -0,0 +1,10 @@
+interface NotifyVideoCompletedParams {
+  generationBatchId: string;
+  model: string;
+  prompt: string;
+  topicId?: string;
+  userId: string;
+}
+
+// eslint-disable-next-line unused-imports/no-unused-vars
+export async function notifyVideoCompleted(params: NotifyVideoCompletedParams): Promise<void> {}
@@ -0,0 +1,25 @@
+import { describe, expect, it } from 'vitest';
+
+import { formatToken } from './TokenProgress';
+
+describe('formatToken', () => {
+  it('should format numbers >= 1M with M suffix', () => {
+    expect(formatToken(1_000_000)).toBe('1M');
+    expect(formatToken(1_500_000)).toBe('1.5M');
+    expect(formatToken(2_000_000)).toBe('2M');
+    expect(formatToken(10_000_000)).toBe('10M');
+  });
+
+  it('should format numbers >= 1K with K suffix', () => {
+    expect(formatToken(1_000)).toBe('1K');
+    expect(formatToken(14_251)).toBe('14.3K');
+    expect(formatToken(985_749)).toBe('985.7K');
+    expect(formatToken(999_999)).toBe('1000K');
+  });
+
+  it('should format numbers < 1K with comma separator', () => {
+    expect(formatToken(0)).toBe('0');
+    expect(formatToken(1)).toBe('1');
+    expect(formatToken(999)).toBe('999');
+  });
+});
@@ -17,7 +17,11 @@ interface TokenProgressProps {
  showTotal?: string;
 }

-const format = (number: number) => numeral(number).format('0,0');
+export const formatToken = (number: number) => {
+  if (number >= 1_000_000) return numeral(number / 1_000_000).format('0.[0]') + 'M';
+  if (number >= 1_000) return numeral(number / 1_000).format('0.[0]') + 'K';
+  return numeral(number).format('0,0');
+};

 const TokenProgress = memo<TokenProgressProps>(({ data, showIcon, showTotal }) => {
  const total = data.reduce((acc, item) => acc + item.value, 0);
@@ -59,7 +63,7 @@ const TokenProgress = memo<TokenProgressProps>(({ data, showIcon, showTotal }) =
              )}
              <div style={{ color: cssVar.colorTextSecondary }}>{item.title}</div>
            </Flexbox>
-            <div style={{ fontWeight: 500 }}>{format(item.value)}</div>
+            <div style={{ fontWeight: 500 }}>{formatToken(item.value)}</div>
          </Flexbox>
        ))}
        {showTotal && (
@@ -67,7 +71,7 @@ const TokenProgress = memo<TokenProgressProps>(({ data, showIcon, showTotal }) =
            <Divider style={{ marginBlock: 8 }} />
            <Flexbox horizontal align={'center'} gap={4} justify={'space-between'}>
              <div style={{ color: cssVar.colorTextSecondary }}>{showTotal}</div>
-              <div style={{ fontWeight: 500 }}>{format(total)}</div>
+              <div style={{ fontWeight: 500 }}>{formatToken(total)}</div>
            </Flexbox>
          </>
        )}
@@ -0,0 +1,57 @@
+import { Icon } from '@lobehub/ui';
+import { Button } from 'antd';
+import { Minimize2 } from 'lucide-react';
+import { memo, useCallback, useState } from 'react';
+import { useTranslation } from 'react-i18next';
+
+import { useChatStore } from '@/store/chat';
+
+import { useConversationStore } from '../store';
+import BaseErrorForm from './BaseErrorForm';
+
+interface ExceededContextWindowErrorProps {
+  id: string;
+}
+
+const ExceededContextWindowError = memo<ExceededContextWindowErrorProps>(({ id }) => {
+  const { t } = useTranslation('error');
+  const [loading, setLoading] = useState(false);
+
+  const context = useConversationStore((s) => s.context);
+  const regenerateUserMessage = useConversationStore((s) => s.regenerateUserMessage);
+  const parentId = useConversationStore(
+    (s) => s.displayMessages.find((m) => m.id === id)?.parentId,
+  );
+
+  const handleCompact = useCallback(async () => {
+    if (!context.topicId || !parentId) return;
+
+    setLoading(true);
+    try {
+      await useChatStore.getState().executeCompression(context, '');
+      await regenerateUserMessage(parentId);
+    } finally {
+      setLoading(false);
+    }
+  }, [context, parentId, regenerateUserMessage]);
+
+  return (
+    <BaseErrorForm
+      avatar={<Icon icon={Minimize2} size={24} />}
+      desc={t('exceededContext.desc')}
+      title={t('exceededContext.title')}
+      action={
+        <Button
+          disabled={!context.topicId}
+          loading={loading}
+          type={'primary'}
+          onClick={handleCompact}
+        >
+          {t('exceededContext.compact')}
+        </Button>
+      }
+    />
+  );
+});
+
+export default ExceededContextWindowError;
@@ -38,6 +38,11 @@ const loading = () => (
  </Block>
 );

+const ExceededContextWindowError = dynamic(() => import('./ExceededContextWindowError'), {
+  loading,
+  ssr: false,
+});
+
 const OllamaBizError = dynamic(() => import('./OllamaBizError'), { loading, ssr: false });

 const OllamaSetupGuide = dynamic(() => import('./OllamaSetupGuide'), {
@@ -130,6 +135,10 @@ const ErrorMessageExtra = memo<ErrorExtraProps>(({ error: alertError, data }) =>
      return <OllamaBizError {...data} />;
    }

+    case AgentRuntimeErrorType.ExceededContextWindow: {
+      return <ExceededContextWindowError id={data.id} />;
+    }
+
    /* ↓ cloud slot ↓ */

    /* ↑ cloud slot ↑ */
@@ -128,15 +128,7 @@ const SideBarHeaderLayout = memo<SideBarHeaderLayoutProps>(
        padding={6}
      >
        {leftContent}
-        <Flexbox
-          horizontal
-          align={'center'}
-          gap={2}
-          justify={'flex-end'}
-          style={{
-            overflow: 'hidden',
-          }}
-        >
+        <Flexbox horizontal align={'center'} gap={2} justify={'flex-end'}>
          {showTogglePanelButton && <ToggleLeftPanelButton />}
          {right}
        </Flexbox>
@@ -102,12 +102,12 @@ const NavItem = memo<NavItemProps>(
        paddingInline={4}
        variant={variant}
        onClick={(e) => {
-          if (disabled || loading) return;
-          // Prevent default link behavior for normal clicks (let onClick handle it)
-          // But allow cmd+click to open in new tab
+          // Always prevent default <a> navigation for normal clicks to avoid full page reload.
+          // This must run before any early return to ensure SPA navigation is never bypassed.
          if (href && !isModifierClick(e)) {
            e.preventDefault();
          }
+          if (disabled || loading) return;
          onClick?.(e);
        }}
        {...linkProps}
@@ -1,40 +0,0 @@
-import { type PluginRender, type PluginRenderProps } from '@lobehub/chat-plugin-sdk/client';
-import { Skeleton } from '@lobehub/ui';
-import { memo, useEffect, useState } from 'react';
-
-import { system } from './utils';
-
-interface SystemJsRenderProps extends PluginRenderProps {
-  url: string;
-}
-
-const RenderCache: {
-  [url: string]: PluginRender;
-} = {};
-
-const SystemJsRender = memo<SystemJsRenderProps>(({ url, ...props }) => {
-  const [component, setComp] = useState<PluginRender | undefined>(RenderCache[url]);
-
-  useEffect(() => {
-    system
-      .import(url)
-      .then((module1) => {
-        setComp(module1.default);
-        RenderCache[url] = module1.default;
-        // Use the module1 module
-      })
-      .catch((error) => {
-        setComp(undefined);
-        console.error(error);
-      });
-  }, [url]);
-
-  if (!component) {
-    return <Skeleton active style={{ width: 300 }} />;
-  }
-
-  const Render = component;
-
-  return <Render {...props} />;
-});
-export default SystemJsRender;
@@ -1,25 +0,0 @@
-/**
- * This dynamic loading module is implemented using SystemJS, caching four modules in Lobe Chat: React, ReactDOM, antd, and antd-style.
- */
-import 'systemjs';
-
-import * as antd from 'antd';
-import * as AntdStyle from 'antd-style';
-import * as React from 'react';
-import * as ReactDOM from 'react-dom';
-
-System.addImportMap({
-  imports: {
-    'React': 'app:React',
-    'ReactDOM': 'app:ReactDOM',
-    'antd': 'app:antd',
-    'antd-style': 'app:antd-style',
-  },
-});
-
-System.set('app:React', { default: React, ...React });
-System.set('app:ReactDOM', { __useDefault: true, ...ReactDOM });
-System.set('app:antd', antd);
-System.set('app:antd-style', AntdStyle);
-
-export const system = System;
@@ -1,7 +1,5 @@
-import { Skeleton } from '@lobehub/ui';
-import { memo,Suspense } from 'react';
+import { memo } from 'react';

-import dynamic from '@/libs/next/dynamic';
 import { useToolStore } from '@/store/tool';
 import { pluginSelectors } from '@/store/tool/selectors';

@@ -9,8 +7,6 @@ import Loading from '../Loading';
 import { useParseContent } from '../useParseContent';
 import IFrameRender from './IFrameRender';

-const SystemJsRender = dynamic(() => import('./SystemJsRender'), { ssr: false });
-
 export interface PluginDefaultTypeProps {
  content: string;
  loading?: boolean;
@@ -32,13 +28,6 @@ const PluginDefaultType = memo<PluginDefaultTypeProps>(({ content, name, loading

  if (!ui.url) return;

-  if (ui.mode === 'module')
-    return (
-      <Suspense fallback={<Skeleton active style={{ width: 400 }} />}>
-        <SystemJsRender content={data} name={name || 'unknown'} url={ui.url} />
-      </Suspense>
-    );
-
  return (
    <IFrameRender
      content={data}
@@ -1,4 +1,4 @@
-export const LANGCHAIN_SUPPORT_TEXT_LIST = [
+export const SUPPORT_TEXT_LIST = [
  'txt',
  'markdown',
  'md',
@@ -3,7 +3,6 @@ import * as fs from 'node:fs';
 import { join } from 'node:path';

 import { CodeLoader } from '../index';
-import longResult from './long.json';

 describe('CodeLoader', () => {
  it('split simple code', async () => {
@@ -15,13 +14,12 @@ helloWorld();`;

    const result = await CodeLoader(jsCode, 'js');

-    expect(result).toEqual([
-      {
-        pageContent:
-          'function helloWorld() {\n  console.log("Hello, World!");\n}\n// Call the function\nhelloWorld();',
-        metadata: { loc: { lines: { from: 1, to: 5 } } },
-      },
-    ]);
+    expect(result).toHaveLength(1);
+    expect(result[0].pageContent).toBe(
+      'function helloWorld() {\n  console.log("Hello, World!");\n}\n// Call the function\nhelloWorld();',
+    );
+    expect(result[0].metadata.loc.lines.from).toBe(1);
+    expect(result[0].metadata.loc.lines.to).toBe(5);
  });

  it('split long', async () => {
@@ -29,6 +27,11 @@ helloWorld();`;

    const result = await CodeLoader(code, 'js');

-    expect(result).toEqual(longResult);
+    // Should split long code into multiple chunks
+    expect(result.length).toBeGreaterThan(1);
+    for (const chunk of result) {
+      expect(chunk.pageContent).toBeTruthy();
+      expect(chunk.metadata.loc.lines).toBeDefined();
+    }
  });
 });
@@ -0,0 +1,6 @@
+import { splitCode, type SupportedLanguage } from '../../splitter';
+import { loaderConfig } from '../config';
+
+export const CodeLoader = async (text: string, language: string) => {
+  return splitCode(text, language as SupportedLanguage, loaderConfig);
+};
@@ -7,13 +7,17 @@ import { expect } from 'vitest';
 import { CsVLoader } from '../index';

 describe('CSVLoader', () => {
-  it('should run', async () => {
+  it('should parse CSV rows into documents', async () => {
    const content = fs.readFileSync(join(__dirname, `./demo.csv`), 'utf8');
-
    const fileBlob = new Blob([Buffer.from(content)]);

    const data = await CsVLoader(fileBlob);

-    expect(data).toMatchSnapshot();
+    expect(data.length).toBe(32);
+    // Check first row structure
+    expect(data[0].metadata.line).toBe(1);
+    expect(data[0].metadata.source).toBe('blob');
+    expect(data[0].pageContent).toContain('Hair:');
+    expect(data[0].pageContent).toContain('Eye:');
  });
 });
@@ -0,0 +1,24 @@
+import { type DocumentChunk } from '../../types';
+
+export const CsVLoader = async (fileBlob: Blob): Promise<DocumentChunk[]> => {
+  const { dsvFormat } = await import('d3-dsv');
+  const csvParse = dsvFormat(',');
+
+  const text = await fileBlob.text();
+  const rows = csvParse.parse(text);
+
+  return rows.map((row, index) => {
+    const content = Object.entries(row)
+      .filter(([key]) => key !== 'columns')
+      .map(([key, value]) => `${key}: ${value}`)
+      .join('\n');
+
+    return {
+      metadata: {
+        line: index + 1,
+        source: 'blob',
+      },
+      pageContent: content,
+    };
+  });
+};
@@ -0,0 +1,15 @@
+import { splitText } from '../../splitter';
+import { type DocumentChunk } from '../../types';
+import { loaderConfig } from '../config';
+
+export const DocxLoader = async (fileBlob: Blob | string): Promise<DocumentChunk[]> => {
+  const mammoth = await import('mammoth');
+
+  const buffer =
+    typeof fileBlob === 'string'
+      ? Buffer.from(fileBlob)
+      : Buffer.from(await fileBlob.arrayBuffer());
+
+  const result = await mammoth.extractRawText({ buffer });
+  return splitText(result.value, loaderConfig);
+};
@@ -6,20 +6,17 @@ import { expect } from 'vitest';

 import { EPubLoader } from '../index';

-function sanitizeDynamicFields(document: any[]) {
-  for (const doc of document) {
-    doc.metadata.source && (doc.metadata.source = '');
-  }
-  return document;
-}
-
 describe('EPubLoader', () => {
-  it('should run', async () => {
+  it('should parse epub content into chunks', async () => {
    const content = fs.readFileSync(join(__dirname, `./demo.epub`));
-
    const fileContent: Uint8Array = new Uint8Array(content);

    const data = await EPubLoader(fileContent);
-    expect(sanitizeDynamicFields(data)).toMatchSnapshot();
+
+    expect(data.length).toBeGreaterThan(0);
+    for (const chunk of data) {
+      expect(chunk.pageContent).toBeTruthy();
+      expect(chunk.metadata).toBeDefined();
+    }
  });
 });
@@ -0,0 +1,52 @@
+import { TempFileManager } from '@/server/utils/tempFileManager';
+import { nanoid } from '@/utils/uuid';
+
+import { splitText } from '../../splitter';
+import { type DocumentChunk } from '../../types';
+import { loaderConfig } from '../config';
+
+export const EPubLoader = async (content: Uint8Array): Promise<DocumentChunk[]> => {
+  const tempManager = new TempFileManager('epub-');
+
+  try {
+    const tempPath = await tempManager.writeTempFile(content, `${nanoid()}.epub`);
+
+    const { EPub } = await import('epub2');
+    const htmlToText = await import('html-to-text');
+
+    const epub = await EPub.createAsync(tempPath);
+    const chapters = epub.flow || [];
+
+    const documents: DocumentChunk[] = [];
+
+    for (const chapter of chapters) {
+      try {
+        const html = await epub.getChapterRawAsync(chapter.id);
+        const text = htmlToText.convert(html, {
+          wordwrap: 80,
+        });
+
+        if (text.trim()) {
+          const chunks = splitText(text, loaderConfig);
+          for (const chunk of chunks) {
+            documents.push({
+              metadata: {
+                ...chunk.metadata,
+                source: tempPath,
+              },
+              pageContent: chunk.pageContent,
+            });
+          }
+        }
+      } catch {
+        // Skip chapters that can't be parsed
+      }
+    }
+
+    return documents;
+  } catch (e) {
+    throw new Error(`EPubLoader error: ${(e as Error).message}`, { cause: e });
+  } finally {
+    tempManager.cleanup();
+  }
+};
@@ -1,9 +1,6 @@
-import { type SupportedTextSplitterLanguage } from 'langchain/text_splitter';
-import { SupportedTextSplitterLanguages } from 'langchain/text_splitter';
-
-import { LANGCHAIN_SUPPORT_TEXT_LIST } from '@/libs/langchain/file';
-import { type LangChainLoaderType } from '@/libs/langchain/types';
-
+import { SUPPORT_TEXT_LIST } from '../file';
+import { SUPPORTED_LANGUAGES, type SupportedLanguage } from '../splitter';
+import { type DocumentChunk, type FileLoaderType } from '../types';
 import { CodeLoader } from './code';
 import { CsVLoader } from './csv';
 import { DocxLoader } from './docx';
@@ -14,15 +11,15 @@ import { PdfLoader } from './pdf';
 import { PPTXLoader } from './pptx';
 import { TextLoader } from './txt';

-class LangChainError extends Error {
+class DocumentLoaderError extends Error {
  constructor(message: string) {
    super(message);
-    this.name = 'LangChainChunkingError';
+    this.name = 'DocumentLoaderError';
  }
 }

 export class ChunkingLoader {
-  partitionContent = async (filename: string, content: Uint8Array) => {
+  partitionContent = async (filename: string, content: Uint8Array): Promise<DocumentChunk[]> => {
    try {
      const fileBlob = new Blob([Buffer.from(content)]);
      const txt = this.uint8ArrayToString(content);
@@ -74,11 +71,11 @@ export class ChunkingLoader {
        }
      }
    } catch (e) {
-      throw new LangChainError((e as Error).message);
+      throw new DocumentLoaderError((e as Error).message);
    }
  };

-  private getType = (filename: string): LangChainLoaderType | undefined => {
+  private getType = (filename: string): FileLoaderType | undefined => {
    if (filename.endsWith('pptx')) {
      return 'ppt';
    }
@@ -109,11 +106,11 @@ export class ChunkingLoader {

    const ext = filename.split('.').pop();

-    if (ext && SupportedTextSplitterLanguages.includes(ext as SupportedTextSplitterLanguage)) {
+    if (ext && SUPPORTED_LANGUAGES.includes(ext as SupportedLanguage)) {
      return 'code';
    }

-    if (ext && LANGCHAIN_SUPPORT_TEXT_LIST.includes(ext)) return 'text';
+    if (ext && SUPPORT_TEXT_LIST.includes(ext)) return 'text';
  };

  private uint8ArrayToString(uint8Array: Uint8Array) {
@@ -7,11 +7,15 @@ import { expect } from 'vitest';
 import { LatexLoader } from '../index';

 describe('LatexLoader', () => {
-  it('should run', async () => {
+  it('should split LaTeX content into chunks', async () => {
    const content = fs.readFileSync(join(__dirname, `./demo.tex`), 'utf8');

    const data = await LatexLoader(content);

-    expect(data).toMatchSnapshot();
+    expect(data.length).toBeGreaterThan(1);
+    for (const chunk of data) {
+      expect(chunk.pageContent).toBeTruthy();
+      expect(chunk.metadata.loc.lines).toBeDefined();
+    }
  });
 });
@@ -0,0 +1,6 @@
+import { splitLatex } from '../../splitter';
+import { loaderConfig } from '../config';
+
+export const LatexLoader = async (text: string) => {
+  return splitLatex(text, loaderConfig);
+};
@@ -5,12 +5,14 @@ import Callout from '@components/markdown/Callout.astro';
 import Section from '@components/markdown/Section.astro';

 # Views (WIP)
+
 <Callout emoji="⚠️" type="warning">
  Views are currently only implemented in the `drizzle-orm`, `drizzle-kit` does not support views yet.
  You can query the views that already exist in the database, but they won't be added to `drizzle-kit` migrations or `db push` as of now.
 </Callout>

 ## Views declaration
+
 There're several ways you can declare views with Drizzle ORM.

 You can declare views that have to be created or you can declare views that already exist in the database.
@@ -21,6 +23,7 @@ When views are created with either inlined or standalone query builders, view co
 yet when you use `sql` you have to explicitly declare view columns schema.

 ### Declaring views
+
 <Tabs items={['PostgreSQL', 'MySQL', 'SQLite']}>
  <Tab>
    <Section>
@@ -40,12 +43,14 @@ yet when you use `sql` you have to explicitly declare view columns schema.
      export const userView = pgView("user_view").as((qb) => qb.select().from(user));
      export const customersView = pgView("customers_view").as((qb) => qb.select().from(user).where(eq(user.role, "customer")));
      ```
+
      ```sql
      CREATE VIEW "user_view" AS SELECT * FROM "user";
      CREATE VIEW "customers_view" AS SELECT * FROM "user" WHERE "role" = 'customer';
      ```
    </Section>
  </Tab>
+
  <Tab>
    <Section>
      ```ts filename="schema.ts" copy {13-14}
@@ -64,12 +69,14 @@ yet when you use `sql` you have to explicitly declare view columns schema.
      export const userView = mysqlView("user_view").as((qb) => qb.select().from(user));
      export const customersView = mysqlView("customers_view").as((qb) => qb.select().from(user).where(eq(user.role, "customer")));
      ```
+
      ```sql
      CREATE VIEW "user_view" AS SELECT * FROM "user";
      CREATE VIEW "customers_view" AS SELECT * FROM "user" WHERE "role" = 'customer';
      ```
    </Section>
  </Tab>
+
  <Tab>
    <Section>
      ```ts filename="schema.ts" copy {13-14}
@@ -88,6 +95,7 @@ yet when you use `sql` you have to explicitly declare view columns schema.
      export const userView = sqliteView("user_view").as((qb) => qb.select().from(user));
      export const customersView = sqliteView("customers_view").as((qb) => qb.select().from(user).where(eq(user.role, "customer")));
      ```
+
      ```sql
      CREATE VIEW "user_view" AS SELECT * FROM "user";
      CREATE VIEW "customers_view" AS SELECT * FROM "user" WHERE "role" = 'customer';
@@ -97,6 +105,7 @@ yet when you use `sql` you have to explicitly declare view columns schema.
 </Tabs>

 If you need a subset of columns you can use `.select({ ... })` method in query builder, like this:
+
 <Section>
  ```ts {4-6}
  export const customersView = pgView("customers_view").as((qb) => {
@@ -109,12 +118,14 @@ If you need a subset of columns you can use `.select({ ... })` method in query b
      .from(user);
  });
  ```
+
  ```sql
  CREATE VIEW "customers_view" AS SELECT "id", "name", "email" FROM "user" WHERE "role" = 'customer';
  ```
 </Section>

 You can also declare views using `standalone query builder`, it works exactly the same way:
+
 <Tabs items={['PostgreSQL', 'MySQL', 'SQLite']}>
  <Tab>
    <Section>
@@ -136,12 +147,14 @@ You can also declare views using `standalone query builder`, it works exactly th
      export const userView = pgView("user_view").as(qb.select().from(user));
      export const customersView = pgView("customers_view").as(qb.select().from(user).where(eq(user.role, "customer")));
      ```
+
      ```sql
      CREATE VIEW "user_view" AS SELECT * FROM "user";
      CREATE VIEW "customers_view" AS SELECT * FROM "user" WHERE "role" = 'customer';
      ```
    </Section>
  </Tab>
+
  <Tab>
    <Section>
      ```ts filename="schema.ts" copy {3, 15-16}
@@ -162,12 +175,14 @@ You can also declare views using `standalone query builder`, it works exactly th
      export const userView = mysqlView("user_view").as(qb.select().from(user));
      export const customersView = mysqlView("customers_view").as(qb.select().from(user).where(eq(user.role, "customer")));
      ```
+
      ```sql
      CREATE VIEW "user_view" AS SELECT * FROM "user";
      CREATE VIEW "customers_view" AS SELECT * FROM "user" WHERE "role" = 'customer';
      ```
    </Section>
  </Tab>
+
  <Tab>
    <Section>
      ```ts filename="schema.ts" copy {3, 15-16}
@@ -188,6 +203,7 @@ You can also declare views using `standalone query builder`, it works exactly th
      export const userView = sqliteView("user_view").as((qb) => qb.select().from(user));
      export const customerView = sqliteView("customers_view").as((qb) => qb.select().from(user).where(eq(user.role, "customer")));
      ```
+
      ```sql
      CREATE VIEW "user_view" AS SELECT * FROM "user";
      CREATE VIEW "customers_view" AS SELECT * FROM "user" WHERE "role" = 'customer';
@@ -197,6 +213,7 @@ You can also declare views using `standalone query builder`, it works exactly th
 </Tabs>

 ### Declaring views with raw SQL
+
 Whenever you need to declare view using a syntax that is not supported by the query builder,
 you can directly use `sql` operator and explicitly specify view columns schema.

@@ -217,8 +234,10 @@ const newYorkers = pgMaterializedView('new_yorkers', {
 ```

 ### Declaring existing views
+
 When you're provided with a read only access to an existing view in the database you should use `.existing()` view configuration,
 `drizzle-kit` will ignore and will not generate a `create view` statement in the generated migration.
+
 ```ts
 export const user = pgTable("user", {
  id: serial("id"),
@@ -246,27 +265,31 @@ export const trimmedUser = pgMaterializedView("trimmed_user", {
 ```

 ### Materialized views
+
 <IsSupportedChipGroup chips={{ 'MySQL': false, 'PostgreSQL': true, 'SQLite': false }} />

 According to the official docs, PostgreSQL has both **[`regular`](https://www.postgresql.org/docs/current/sql-createview.html)**
 and **[`materialized`](https://www.postgresql.org/docs/current/sql-creatematerializedview.html)** views.

 Materialized views in PostgreSQL use the rule system like views do, but persist the results in a table-like form.
+
 {/* This means that when a query is executed against a materialized view, the results are returned directly from the materialized view,
-like from a table, rather than being reconstructed by executing the query against the underlying base tables that make up the view. */}
+  like from a table, rather than being reconstructed by executing the query against the underlying base tables that make up the view. */}

 Drizzle ORM natively supports PostgreSQL materialized views:

 <Section>
-```ts filename="schema.ts" copy
-const newYorkers = pgMaterializedView('new_yorkers').as((qb) => qb.select().from(users).where(eq(users.cityId, 1)));
-```
-```sql
-CREATE MATERIALIZED VIEW "new_yorkers" AS SELECT * FROM "users";
-```
+  ```ts filename="schema.ts" copy
+  const newYorkers = pgMaterializedView('new_yorkers').as((qb) => qb.select().from(users).where(eq(users.cityId, 1)));
+  ```
+
+  ```sql
+  CREATE MATERIALIZED VIEW "new_yorkers" AS SELECT * FROM "users";
+  ```
 </Section>

 You can then refresh materialized views in the application runtime:
+
 ```ts copy
 await db.refreshMaterializedView(newYorkers);

@@ -276,8 +299,9 @@ await db.refreshMaterializedView(newYorkers).withNoData();
 ```

 ### Extended example
+
 <Callout emoji="ℹ️" type="info">
-All the parameters inside the query will be inlined, instead of replaced by `$1`, `$2`, etc.
+  All the parameters inside the query will be inlined, instead of replaced by `$1`, `$2`, etc.
 </Callout>

 ```ts copy
@@ -8,6 +8,12 @@ describe('MarkdownLoader', () => {
  it('should run', async () => {
    const content = fs.readFileSync(join(__dirname, `./demo.mdx`), 'utf8');

-    await MarkdownLoader(content);
+    const result = await MarkdownLoader(content);
+
+    expect(result.length).toBeGreaterThan(0);
+    for (const chunk of result) {
+      expect(chunk.pageContent).toBeTruthy();
+      expect(chunk.metadata.loc.lines).toBeDefined();
+    }
  });
 });
@@ -0,0 +1,6 @@
+import { splitMarkdown } from '../../splitter';
+import { loaderConfig } from '../config';
+
+export const MarkdownLoader = async (text: string) => {
+  return splitMarkdown(text, loaderConfig);
+};
@@ -0,0 +1,20 @@
+import { type DocumentChunk } from '../../types';
+
+export const PdfLoader = async (fileBlob: Blob): Promise<DocumentChunk[]> => {
+  const pdfParse = (await import('pdf-parse')).default;
+
+  const buffer = Buffer.from(await fileBlob.arrayBuffer());
+  const data = await pdfParse(buffer);
+
+  // Split by pages using form feed character, or treat as single page
+  const pages: string[] = data.text
+    ? data.text.split(/\f/).filter((page: string) => page.trim().length > 0)
+    : [];
+
+  return pages.map((pageContent: string, index: number) => ({
+    metadata: {
+      loc: { pageNumber: index + 1 },
+    },
+    pageContent: pageContent.trim(),
+  }));
+};
@@ -0,0 +1,19 @@
+import { type DocumentChunk } from '../../types';
+
+export const PPTXLoader = async (fileBlob: Blob | string): Promise<DocumentChunk[]> => {
+  const { parseOfficeAsync } = await import('officeparser');
+
+  const buffer =
+    typeof fileBlob === 'string'
+      ? Buffer.from(fileBlob)
+      : Buffer.from(await fileBlob.arrayBuffer());
+
+  const text = await parseOfficeAsync(buffer);
+
+  return [
+    {
+      metadata: {},
+      pageContent: text,
+    },
+  ];
+};
@@ -3,7 +3,6 @@ import * as fs from 'node:fs';
 import { join } from 'node:path';

 import { TextLoader } from '../index';
-import longResult from './long.json';

 describe('TextLoader', () => {
  it('split simple content', async () => {
@@ -35,13 +34,11 @@ describe('TextLoader', () => {

    const result = await TextLoader(content);

-    expect(result).toEqual([
-      {
-        pageContent:
-          '好的,我们以基于 Puppeteer 的截图服务为例,给出一个具体的示例:\n\n| 服务器配置 | 并发量 |\n| --- | --- |\n| 1c1g | 50-100 |\n| 2c4g | 200-500 |\n| 4c8g | 500-1000 |\n| 8c16g | 1000-2000 |\n\n这里的并发量是根据以下假设条件估算的:\n\n1. 应用程序使用 Puppeteer 进行网页截图,每个请求需要 500ms-1s 的处理时间。\n2. CPU 密集型任务,CPU 是主要的性能瓶颈。\n3. 每个请求需要 50-100MB 的内存。\n4. 没有其他依赖服务,如数据库等。\n5. 网络带宽足够,不是瓶颈。\n\n在这种情况下:\n\n- 1c1g 的服务器,由于 CPU 资源较少,并发量较低,大约在 50-100 左右。\n- 2c4g 的服务器,CPU 资源增加,并发量可以提高到 200-500 左右。\n- 4c8g 的服务器,CPU 资源进一步增加,并发量可以提高到 500-1000 左右。\n- 8c16g 的服务器,CPU 资源进一步增加,并发量可以提高到 1000-2000 左右。\n\n需要注意的是,这只是一个大致的估计,实际情况可能会有差异。在正式部署时,建议进行负载测试,根据实际情况进行调整和优化。',
-        metadata: { loc: { lines: { from: 1, to: 25 } } },
-      },
-    ]);
+    // Should produce a single chunk for short content
+    expect(result).toHaveLength(1);
+    expect(result[0].pageContent).toBe(content);
+    expect(result[0].metadata.loc.lines.from).toBe(1);
+    expect(result[0].metadata.loc.lines.to).toBe(25);
  });

  it('split long', async () => {
@@ -49,6 +46,13 @@ describe('TextLoader', () => {

    const result = await TextLoader(content);

-    expect(result).toEqual(longResult);
+    // Should split long content into multiple chunks
+    expect(result.length).toBeGreaterThan(1);
+    // Each chunk should have pageContent and metadata
+    for (const chunk of result) {
+      expect(chunk.pageContent).toBeTruthy();
+      expect(chunk.metadata.loc.lines.from).toBeGreaterThanOrEqual(1);
+      expect(chunk.metadata.loc.lines.to).toBeGreaterThanOrEqual(chunk.metadata.loc.lines.from);
+    }
  });
 });
@@ -0,0 +1,6 @@
+import { splitText } from '../../splitter';
+import { loaderConfig } from '../config';
+
+export const TextLoader = async (text: string) => {
+  return splitText(text, loaderConfig);
+};
@@ -0,0 +1,193 @@
+import { type DocumentChunk } from '../types';
+import {
+  DEFAULT_SEPARATORS,
+  getSeparatorsForLanguage,
+  LATEX_SEPARATORS,
+  MARKDOWN_SEPARATORS,
+  type SupportedLanguage,
+} from './separators';
+
+export { SUPPORTED_LANGUAGES, type SupportedLanguage } from './separators';
+
+interface SplitterConfig {
+  chunkOverlap: number;
+  chunkSize: number;
+}
+
+/**
+ * Splits text into overlapping chunks using a recursive separator strategy.
+ * Replicates LangChain's RecursiveCharacterTextSplitter algorithm.
+ */
+function splitTextWithSeparators(
+  text: string,
+  separators: string[],
+  config: SplitterConfig,
+): string[] {
+  const { chunkSize, chunkOverlap } = config;
+
+  // Find the appropriate separator
+  let separator = separators.at(-1)!;
+  let newSeparators: string[] | undefined;
+
+  for (let i = 0; i < separators.length; i++) {
+    const sep = separators[i];
+    if (sep === '') {
+      separator = '';
+      break;
+    }
+    if (text.includes(sep)) {
+      separator = sep;
+      newSeparators = separators.slice(i + 1);
+      break;
+    }
+  }
+
+  // Split the text by the chosen separator
+  const splits = separator ? text.split(separator) : [...text];
+
+  // Merge splits into chunks respecting chunkSize
+  const goodSplits: string[] = [];
+  const finalChunks: string[] = [];
+
+  for (const s of splits) {
+    if (s.length < chunkSize) {
+      goodSplits.push(s);
+    } else {
+      if (goodSplits.length > 0) {
+        const merged = mergeSplits(goodSplits, separator, config);
+        finalChunks.push(...merged);
+        goodSplits.length = 0;
+      }
+      // If this piece is still too large and we have more separators, recurse
+      if (newSeparators && newSeparators.length > 0) {
+        const subChunks = splitTextWithSeparators(s, newSeparators, config);
+        finalChunks.push(...subChunks);
+      } else {
+        finalChunks.push(s);
+      }
+    }
+  }
+
+  if (goodSplits.length > 0) {
+    const merged = mergeSplits(goodSplits, separator, config);
+    finalChunks.push(...merged);
+  }
+
+  return finalChunks;
+}
+
+/**
+ * Merge small splits into chunks respecting chunkSize and chunkOverlap.
+ */
+function mergeSplits(splits: string[], separator: string, config: SplitterConfig): string[] {
+  const { chunkSize, chunkOverlap } = config;
+  const chunks: string[] = [];
+  const currentChunk: string[] = [];
+  let total = 0;
+
+  for (const s of splits) {
+    const len = s.length;
+    const sepLen = currentChunk.length > 0 ? separator.length : 0;
+
+    if (total + len + sepLen > chunkSize && currentChunk.length > 0) {
+      const chunk = currentChunk.join(separator);
+      if (chunk.length > 0) {
+        chunks.push(chunk);
+      }
+
+      // Keep overlap: drop from the start of currentChunk until we fit in overlap
+      while (total > chunkOverlap || (total + len + separator.length > chunkSize && total > 0)) {
+        if (currentChunk.length === 0) break;
+        const removed = currentChunk.shift()!;
+        total -= removed.length + (currentChunk.length > 0 ? separator.length : 0);
+      }
+    }
+
+    currentChunk.push(s);
+    total += len + (currentChunk.length > 1 ? separator.length : 0);
+  }
+
+  const lastChunk = currentChunk.join(separator);
+  if (lastChunk.length > 0) {
+    chunks.push(lastChunk);
+  }
+
+  return chunks;
+}
+
+/**
+ * Calculate line location metadata for a chunk within the original text.
+ */
+function getLineLocation(fullText: string, chunk: string): { from: number; to: number } {
+  const index = fullText.indexOf(chunk);
+  if (index === -1) {
+    return { from: 1, to: 1 };
+  }
+
+  const beforeChunk = fullText.slice(0, index);
+  const from = beforeChunk.split('\n').length;
+  const chunkLines = chunk.split('\n').length;
+  const to = from + chunkLines - 1;
+
+  return { from, to };
+}
+
+/**
+ * Create document chunks from text using given separators.
+ */
+function createDocuments(
+  text: string,
+  separators: string[],
+  config: SplitterConfig,
+  baseMetadata?: Record<string, any>,
+): DocumentChunk[] {
+  const chunks = splitTextWithSeparators(text, separators, config);
+
+  // Track search position to handle duplicate chunks correctly
+  let searchFrom = 0;
+
+  return chunks.map((chunk) => {
+    const index = text.indexOf(chunk, searchFrom);
+    let loc = { from: 1, to: 1 };
+
+    if (index !== -1) {
+      const beforeChunk = text.slice(0, index);
+      const from = beforeChunk.split('\n').length;
+      const chunkLines = chunk.split('\n').length;
+      loc = { from, to: from + chunkLines - 1 };
+      // Advance search position past this match (but allow overlap)
+      searchFrom = index + 1;
+    }
+
+    return {
+      metadata: {
+        ...baseMetadata,
+        loc: { lines: loc },
+      },
+      pageContent: chunk,
+    };
+  });
+}
+
+// --- Public API ---
+
+export function splitText(text: string, config: SplitterConfig): DocumentChunk[] {
+  return createDocuments(text, DEFAULT_SEPARATORS, config);
+}
+
+export function splitMarkdown(text: string, config: SplitterConfig): DocumentChunk[] {
+  return createDocuments(text, MARKDOWN_SEPARATORS, config);
+}
+
+export function splitLatex(text: string, config: SplitterConfig): DocumentChunk[] {
+  return createDocuments(text, LATEX_SEPARATORS, config);
+}
+
+export function splitCode(
+  text: string,
+  language: SupportedLanguage,
+  config: SplitterConfig,
+): DocumentChunk[] {
+  const separators = getSeparatorsForLanguage(language);
+  return createDocuments(text, separators, config);
+}
@@ -0,0 +1,297 @@
+/**
+ * Language-specific separators for recursive text splitting.
+ * Each array is ordered from most to least specific separator.
+ */
+
+export type SupportedLanguage =
+  | 'cpp'
+  | 'go'
+  | 'java'
+  | 'js'
+  | 'php'
+  | 'proto'
+  | 'python'
+  | 'rst'
+  | 'ruby'
+  | 'rust'
+  | 'scala'
+  | 'swift'
+  | 'markdown'
+  | 'latex'
+  | 'html'
+  | 'sol';
+
+export const SUPPORTED_LANGUAGES: SupportedLanguage[] = [
+  'cpp',
+  'go',
+  'java',
+  'js',
+  'php',
+  'proto',
+  'python',
+  'rst',
+  'ruby',
+  'rust',
+  'scala',
+  'swift',
+  'markdown',
+  'latex',
+  'html',
+  'sol',
+];
+
+export const DEFAULT_SEPARATORS = ['\n\n', '\n', ' ', ''];
+
+export const MARKDOWN_SEPARATORS = [
+  '\n## ',
+  '\n### ',
+  '\n#### ',
+  '\n##### ',
+  '\n###### ',
+  '```\n\n',
+  '\n\n***\n\n',
+  '\n\n---\n\n',
+  '\n\n___\n\n',
+  '\n\n',
+  '\n',
+  ' ',
+  '',
+];
+
+export const LATEX_SEPARATORS = [
+  '\n\\chapter{',
+  '\n\\section{',
+  '\n\\subsection{',
+  '\n\\subsubsection{',
+  '\n\\begin{enumerate}',
+  '\n\\begin{itemize}',
+  '\n\\begin{description}',
+  '\n\\begin{list}',
+  '\n\\begin{quote}',
+  '\n\\begin{quotation}',
+  '\n\\begin{verse}',
+  '\n\\begin{verbatim}',
+  '\n\\begin{align}',
+  '$$',
+  '$',
+  '\n\n',
+  '\n',
+  ' ',
+  '',
+];
+
+const LANGUAGE_SEPARATORS: Record<SupportedLanguage, string[]> = {
+  cpp: [
+    '\nclass ',
+    '\nvoid ',
+    '\nint ',
+    '\nfloat ',
+    '\ndouble ',
+    '\nif ',
+    '\nfor ',
+    '\nwhile ',
+    '\nswitch ',
+    '\ncase ',
+    '\n\n',
+    '\n',
+    ' ',
+    '',
+  ],
+  go: [
+    '\nfunc ',
+    '\nvar ',
+    '\nconst ',
+    '\ntype ',
+    '\nif ',
+    '\nfor ',
+    '\nswitch ',
+    '\ncase ',
+    '\n\n',
+    '\n',
+    ' ',
+    '',
+  ],
+  html: [
+    '<body>',
+    '<div>',
+    '<p>',
+    '<br>',
+    '<li>',
+    '<h1>',
+    '<h2>',
+    '<h3>',
+    '<h4>',
+    '<h5>',
+    '<h6>',
+    '<span>',
+    '<table>',
+    '<tr>',
+    '<td>',
+    '<th>',
+    '<ul>',
+    '<ol>',
+    '<header>',
+    '<footer>',
+    '<nav>',
+    '<head>',
+    '<style>',
+    '<script>',
+    '<meta>',
+    '<title>',
+    ' ',
+    '',
+  ],
+  java: [
+    '\nclass ',
+    '\npublic ',
+    '\nprotected ',
+    '\nprivate ',
+    '\nstatic ',
+    '\nif ',
+    '\nfor ',
+    '\nwhile ',
+    '\nswitch ',
+    '\ncase ',
+    '\n\n',
+    '\n',
+    ' ',
+    '',
+  ],
+  js: [
+    '\nfunction ',
+    '\nconst ',
+    '\nlet ',
+    '\nvar ',
+    '\nclass ',
+    '\nif ',
+    '\nfor ',
+    '\nwhile ',
+    '\nswitch ',
+    '\ncase ',
+    '\ndefault ',
+    '\n\n',
+    '\n',
+    ' ',
+    '',
+  ],
+  latex: LATEX_SEPARATORS,
+  markdown: MARKDOWN_SEPARATORS,
+  php: [
+    '\nfunction ',
+    '\nclass ',
+    '\nif ',
+    '\nforeach ',
+    '\nwhile ',
+    '\ndo ',
+    '\nswitch ',
+    '\ncase ',
+    '\n\n',
+    '\n',
+    ' ',
+    '',
+  ],
+  proto: [
+    '\nmessage ',
+    '\nservice ',
+    '\nenum ',
+    '\noption ',
+    '\nimport ',
+    '\nsyntax ',
+    '\n\n',
+    '\n',
+    ' ',
+    '',
+  ],
+  python: ['\nclass ', '\ndef ', '\n\tdef ', '\n\n', '\n', ' ', ''],
+  rst: ['\n===\n', '\n---\n', '\n***\n', '\n.. ', '\n\n', '\n', ' ', ''],
+  ruby: [
+    '\ndef ',
+    '\nclass ',
+    '\nif ',
+    '\nunless ',
+    '\nwhile ',
+    '\nfor ',
+    '\ndo ',
+    '\nbegin ',
+    '\nrescue ',
+    '\n\n',
+    '\n',
+    ' ',
+    '',
+  ],
+  rust: [
+    '\nfn ',
+    '\nconst ',
+    '\nlet ',
+    '\nif ',
+    '\nwhile ',
+    '\nfor ',
+    '\nloop ',
+    '\nmatch ',
+    '\nconst ',
+    '\n\n',
+    '\n',
+    ' ',
+    '',
+  ],
+  scala: [
+    '\nclass ',
+    '\nobject ',
+    '\ndef ',
+    '\nval ',
+    '\nvar ',
+    '\nif ',
+    '\nfor ',
+    '\nwhile ',
+    '\nmatch ',
+    '\ncase ',
+    '\n\n',
+    '\n',
+    ' ',
+    '',
+  ],
+  sol: [
+    '\npragma ',
+    '\nusing ',
+    '\ncontract ',
+    '\ninterface ',
+    '\nlibrary ',
+    '\nconstructor ',
+    '\ntype ',
+    '\nfunction ',
+    '\nevent ',
+    '\nmodifier ',
+    '\nerror ',
+    '\nstruct ',
+    '\nenum ',
+    '\nif ',
+    '\nfor ',
+    '\nwhile ',
+    '\ndo while ',
+    '\nassembly ',
+    '\n\n',
+    '\n',
+    ' ',
+    '',
+  ],
+  swift: [
+    '\nfunc ',
+    '\nclass ',
+    '\nstruct ',
+    '\nenum ',
+    '\nif ',
+    '\nfor ',
+    '\nwhile ',
+    '\ndo ',
+    '\nswitch ',
+    '\ncase ',
+    '\n\n',
+    '\n',
+    ' ',
+    '',
+  ],
+};
+
+export function getSeparatorsForLanguage(language: SupportedLanguage): string[] {
+  return LANGUAGE_SEPARATORS[language];
+}
@@ -0,0 +1,16 @@
+export interface DocumentChunk {
+  id?: string;
+  metadata: Record<string, any>;
+  pageContent: string;
+}
+
+export type FileLoaderType =
+  | 'code'
+  | 'ppt'
+  | 'pdf'
+  | 'markdown'
+  | 'doc'
+  | 'text'
+  | 'latex'
+  | 'csv'
+  | 'epub';
@@ -1,13 +0,0 @@
-import { type SupportedTextSplitterLanguage } from 'langchain/text_splitter';
-import { RecursiveCharacterTextSplitter } from 'langchain/text_splitter';
-
-import { loaderConfig } from '@/libs/langchain/loaders/config';
-
-export const CodeLoader = async (text: string, language: string) => {
-  const splitter = RecursiveCharacterTextSplitter.fromLanguage(
-    language as SupportedTextSplitterLanguage,
-    loaderConfig,
-  );
-
-  return await splitter.createDocuments([text]);
-};
@@ -1,422 +0,0 @@
-// Vitest Snapshot v1, https://vitest.dev/guide/snapshot.html
-
-exports[`CSVLoader > should run 1`] = `
-[
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 1,
-      "source": "blob",
-    },
-    "pageContent": ": 1
-Hair: Black
-Eye: Brown
-Sex: Male
-Freq: 32",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 2,
-      "source": "blob",
-    },
-    "pageContent": ": 2
-Hair: Brown
-Eye: Brown
-Sex: Male
-Freq: 53",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 3,
-      "source": "blob",
-    },
-    "pageContent": ": 3
-Hair: Red
-Eye: Brown
-Sex: Male
-Freq: 10",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 4,
-      "source": "blob",
-    },
-    "pageContent": ": 4
-Hair: Blond
-Eye: Brown
-Sex: Male
-Freq: 3",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 5,
-      "source": "blob",
-    },
-    "pageContent": ": 5
-Hair: Black
-Eye: Blue
-Sex: Male
-Freq: 11",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 6,
-      "source": "blob",
-    },
-    "pageContent": ": 6
-Hair: Brown
-Eye: Blue
-Sex: Male
-Freq: 50",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 7,
-      "source": "blob",
-    },
-    "pageContent": ": 7
-Hair: Red
-Eye: Blue
-Sex: Male
-Freq: 10",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 8,
-      "source": "blob",
-    },
-    "pageContent": ": 8
-Hair: Blond
-Eye: Blue
-Sex: Male
-Freq: 30",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 9,
-      "source": "blob",
-    },
-    "pageContent": ": 9
-Hair: Black
-Eye: Hazel
-Sex: Male
-Freq: 10",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 10,
-      "source": "blob",
-    },
-    "pageContent": ": 10
-Hair: Brown
-Eye: Hazel
-Sex: Male
-Freq: 25",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 11,
-      "source": "blob",
-    },
-    "pageContent": ": 11
-Hair: Red
-Eye: Hazel
-Sex: Male
-Freq: 7",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 12,
-      "source": "blob",
-    },
-    "pageContent": ": 12
-Hair: Blond
-Eye: Hazel
-Sex: Male
-Freq: 5",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 13,
-      "source": "blob",
-    },
-    "pageContent": ": 13
-Hair: Black
-Eye: Green
-Sex: Male
-Freq: 3",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 14,
-      "source": "blob",
-    },
-    "pageContent": ": 14
-Hair: Brown
-Eye: Green
-Sex: Male
-Freq: 15",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 15,
-      "source": "blob",
-    },
-    "pageContent": ": 15
-Hair: Red
-Eye: Green
-Sex: Male
-Freq: 7",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 16,
-      "source": "blob",
-    },
-    "pageContent": ": 16
-Hair: Blond
-Eye: Green
-Sex: Male
-Freq: 8",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 17,
-      "source": "blob",
-    },
-    "pageContent": ": 17
-Hair: Black
-Eye: Brown
-Sex: Female
-Freq: 36",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 18,
-      "source": "blob",
-    },
-    "pageContent": ": 18
-Hair: Brown
-Eye: Brown
-Sex: Female
-Freq: 66",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 19,
-      "source": "blob",
-    },
-    "pageContent": ": 19
-Hair: Red
-Eye: Brown
-Sex: Female
-Freq: 16",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 20,
-      "source": "blob",
-    },
-    "pageContent": ": 20
-Hair: Blond
-Eye: Brown
-Sex: Female
-Freq: 4",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 21,
-      "source": "blob",
-    },
-    "pageContent": ": 21
-Hair: Black
-Eye: Blue
-Sex: Female
-Freq: 9",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 22,
-      "source": "blob",
-    },
-    "pageContent": ": 22
-Hair: Brown
-Eye: Blue
-Sex: Female
-Freq: 34",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 23,
-      "source": "blob",
-    },
-    "pageContent": ": 23
-Hair: Red
-Eye: Blue
-Sex: Female
-Freq: 7",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 24,
-      "source": "blob",
-    },
-    "pageContent": ": 24
-Hair: Blond
-Eye: Blue
-Sex: Female
-Freq: 64",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 25,
-      "source": "blob",
-    },
-    "pageContent": ": 25
-Hair: Black
-Eye: Hazel
-Sex: Female
-Freq: 5",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 26,
-      "source": "blob",
-    },
-    "pageContent": ": 26
-Hair: Brown
-Eye: Hazel
-Sex: Female
-Freq: 29",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 27,
-      "source": "blob",
-    },
-    "pageContent": ": 27
-Hair: Red
-Eye: Hazel
-Sex: Female
-Freq: 7",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 28,
-      "source": "blob",
-    },
-    "pageContent": ": 28
-Hair: Blond
-Eye: Hazel
-Sex: Female
-Freq: 5",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 29,
-      "source": "blob",
-    },
-    "pageContent": ": 29
-Hair: Black
-Eye: Green
-Sex: Female
-Freq: 2",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 30,
-      "source": "blob",
-    },
-    "pageContent": ": 30
-Hair: Brown
-Eye: Green
-Sex: Female
-Freq: 14",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 31,
-      "source": "blob",
-    },
-    "pageContent": ": 31
-Hair: Red
-Eye: Green
-Sex: Female
-Freq: 7",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "blobType": "",
-      "line": 32,
-      "source": "blob",
-    },
-    "pageContent": ": 32
-Hair: Blond
-Eye: Green
-Sex: Female
-Freq: 8",
-  },
-]
-`;
@@ -1,7 +0,0 @@
-import { CSVLoader } from '@langchain/community/document_loaders/fs/csv';
-
-export const CsVLoader = async (fileBlob: Blob) => {
-  const loader = new CSVLoader(fileBlob);
-
-  return await loader.load();
-};
@@ -1,13 +0,0 @@
-import { DocxLoader as Loader } from '@langchain/community/document_loaders/fs/docx';
-import { RecursiveCharacterTextSplitter } from 'langchain/text_splitter';
-
-import { loaderConfig } from '../config';
-
-export const DocxLoader = async (fileBlob: Blob | string) => {
-  const loader = new Loader(fileBlob);
-
-  const splitter = new RecursiveCharacterTextSplitter(loaderConfig);
-  const documents = await loader.load();
-
-  return await splitter.splitDocuments(documents);
-};
@@ -1,238 +0,0 @@
-// Vitest Snapshot v1, https://vitest.dev/guide/snapshot.html
-
-exports[`EPubLoader > should run 1`] = `
-[
-  Document {
-    "id": undefined,
-    "metadata": {
-      "loc": {
-        "lines": {
-          "from": 1,
-          "to": 13,
-        },
-      },
-      "source": "",
-    },
-    "pageContent": "HEFTY WATER
-
-This document serves to test Reading System support for the epub:switch
-[http://idpf.org/epub/30/spec/epub30-contentdocs.html#sec-xhtml-content-switch]
-element. There is also a little bit of ruby markup
-[http://www.w3.org/TR/html5/the-ruby-element.html#the-ruby-element] available.
-
-
-THE SWITCH
-
-Below is an instance of the epub:switch element, containing Chemical Markup
-Language [http://en.wikipedia.org/wiki/Chemical_Markup_Language] (CML). The
-fallback content is a chunk of plain XHTML5.",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "loc": {
-        "lines": {
-          "from": 9,
-          "to": 22,
-        },
-      },
-      "source": "",
-    },
-    "pageContent": "THE SWITCH
-
-Below is an instance of the epub:switch element, containing Chemical Markup
-Language [http://en.wikipedia.org/wiki/Chemical_Markup_Language] (CML). The
-fallback content is a chunk of plain XHTML5.
-
- * If your Reading System supports epub:switch and CML, it will render the CML
-   formula natively, and ignore (a.k.a not display) the XHTML fallback.
- * If your Reading System supports epub:switch but not CML, it will ignore (not
-   display) the CML formula, and render the the XHTML fallback instead.
- * If your Reading System does not support epub:switch at all, then the
-   rendering results are somewhat unpredictable, but the most likely result is
-   that it will display both a failed attempt to render the CML and the XHTML
-   fallback.",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "loc": {
-        "lines": {
-          "from": 24,
-          "to": 43,
-        },
-      },
-      "source": "",
-    },
-    "pageContent": "Note: the XHTML fallback is bold and enclosed in a gray dotted box with a
-slightly gray background. A failed CML rendering will most likely appear above
-the gray fallback box and read:
-"H hydrogen O oxygen hefty H O water".
-
-Here the switch begins...
-
-
-H hydrogen O oxygen hefty H O water
-
-2H2 + O2 ⟶ 2H2O
-
-... and here the switch ends.
-
-
-THE SOURCE
-
-Below is a rendition of the source code of the switch element. Your Reading
-System should display this correctly regardless of whether it supports the
-switch element.",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "loc": {
-        "lines": {
-          "from": 46,
-          "to": 66,
-        },
-      },
-      "source": "",
-    },
-    "pageContent": "<switch xmlns="http://www.idpf.org/2007/ops">
-    <case required-namespace="http://www.xml-cml.org/schema">
-        <chem xmlns="http://www.xml-cml.org/schema">
-            <reaction>
-                <molecule n="2">
-                    <atom n="2"> H </atom>
-                    <caption> hydrogen </caption>
-                </molecule>
-                <plus></plus>
-                <molecule>
-                    <atom n="2"> O </atom>
-                    <caption> oxygen </caption>
-                </molecule>
-                <gives>
-                    <caption> hefty </caption>
-                </gives>
-                <molecule n="2">
-                    <atom n="2"> H </atom>
-                    <atom> O </atom>
-                    <caption> water </caption>
-                </molecule>",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "loc": {
-        "lines": {
-          "from": 57,
-          "to": 79,
-        },
-      },
-      "source": "",
-    },
-    "pageContent": "<caption> oxygen </caption>
-                </molecule>
-                <gives>
-                    <caption> hefty </caption>
-                </gives>
-                <molecule n="2">
-                    <atom n="2"> H </atom>
-                    <atom> O </atom>
-                    <caption> water </caption>
-                </molecule>
-            </reaction>
-        </chem>
-    </case>
-    <default>
-        <p xmlns="http://www.w3.org/1999/xhtml" id="fallback">
-            <span>2H<sub>2</sub></span>
-            <span>+</span>
-            <span>O<sub>2</sub></span>
-            <span>⟶</span>
-            <span>2H<sub>2</sub>O</span>
-        </p>
-    </default>                
-</switch>",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "loc": {
-        "lines": {
-          "from": 84,
-          "to": 94,
-        },
-      },
-      "source": "",
-    },
-    "pageContent": "HEFTY RUBY WATER
-
-While the ruby element is mostly used in east-asian languages, it can also be
-useful in other contexts. As an example, and as you can see in the source of the
-CML element above, the code includes a caption element which is intended to be
-displayed below the formula segments. Following this paragraph is a reworked
-version of the XHTML fallback used above, using the ruby element. If your
-Reading System does not support ruby markup, then the captions will appear in
-parentheses on the same line as the formula segments.
-
-2H2(hydrogen) + O2(oxygen) ⟶(hefty) 2H2O(water)",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "loc": {
-        "lines": {
-          "from": 94,
-          "to": 111,
-        },
-      },
-      "source": "",
-    },
-    "pageContent": "2H2(hydrogen) + O2(oxygen) ⟶(hefty) 2H2O(water)
-
-If your Reading System in addition to supporting ruby markup also supports the
-epub-ruby-position
-[http://idpf.org/epub/30/spec/epub30-contentdocs.html#sec-css-ruby-position]
-property, then the captions will appear under the formula segments instead of
-over them.
-
-The source code for the ruby version of the XHTML fallback looks as follows:
-
-
-<p id="rubyp">
-    <ruby>2H<sub>2</sub><rp>(</rp><rt>hydrogen</rt><rp>)</rp></ruby>
-    <span>+</span>
-    <ruby>O<sub>2</sub><rp>(</rp><rt>oxygen</rt><rp>)</rp></ruby>
-    <ruby>⟶<rp>(</rp><rt>hefty</rt><rp>)</rp></ruby>
-    <ruby>2H<sub>2</sub>O<rp>(</rp><rt>water</rt><rp>)</rp></ruby>
-</p>",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "loc": {
-        "lines": {
-          "from": 105,
-          "to": 120,
-        },
-      },
-      "source": "",
-    },
-    "pageContent": "<p id="rubyp">
-    <ruby>2H<sub>2</sub><rp>(</rp><rt>hydrogen</rt><rp>)</rp></ruby>
-    <span>+</span>
-    <ruby>O<sub>2</sub><rp>(</rp><rt>oxygen</rt><rp>)</rp></ruby>
-    <ruby>⟶<rp>(</rp><rt>hefty</rt><rp>)</rp></ruby>
-    <ruby>2H<sub>2</sub>O<rp>(</rp><rt>water</rt><rp>)</rp></ruby>
-</p>                
-            
-
-... and the css declaration using the -epub-ruby-position property looks like
-this:
-
-
-p#rubyp {
-    -epub-ruby-position : under;
-}",
-  },
-]
-`;
@@ -1,24 +0,0 @@
-import { EPubLoader as Loader } from '@langchain/community/document_loaders/fs/epub';
-import { RecursiveCharacterTextSplitter } from 'langchain/text_splitter';
-
-import { TempFileManager } from '@/server/utils/tempFileManager';
-import { nanoid } from '@/utils/uuid';
-
-import { loaderConfig } from '../config';
-
-export const EPubLoader = async (content: Uint8Array) => {
-  const tempManager = new TempFileManager('epub-');
-
-  try {
-    const tempPath = await tempManager.writeTempFile(content, `${nanoid()}.epub`);
-    const loader = new Loader(tempPath);
-    const documents = await loader.load();
-
-    const splitter = new RecursiveCharacterTextSplitter(loaderConfig);
-    return await splitter.splitDocuments(documents);
-  } catch (e) {
-    throw new Error(`EPubLoader error: ${(e as Error).message}`, { cause: e });
-  } finally {
-    tempManager.cleanup(); // Ensure cleanup
-  }
-};
@@ -1,205 +0,0 @@
-// Vitest Snapshot v1, https://vitest.dev/guide/snapshot.html
-
-exports[`LatexLoader > should run 1`] = `
-[
-  Document {
-    "id": undefined,
-    "metadata": {
-      "loc": {
-        "lines": {
-          "from": 1,
-          "to": 41,
-        },
-      },
-    },
-    "pageContent": "\\documentclass{article}
-
-
-\\usepackage{graphicx} % Required for inserting images
-\\usepackage{amsmath}  % Required for mathematical symbols
-\\usepackage{hyperref} % For hyperlinks
-
-
-\\title{Sample LaTeX Document}
-\\author{Generated by ChatGPT}
-\\date{\\today}
-
-
-\\begin{document}
-
-
-\\maketitle
-
-
-\\tableofcontents
-
-
-\\section{Introduction}
-This is a sample LaTeX document that includes various common elements such as sections, lists, tables, figures, and mathematical equations.
-
-
-\\section{Lists}
-\\subsection{Itemized List}
-\\begin{itemize}
-\\item First item
-\\item Second item
-\\item Third item
-\\end{itemize}
-
-
-\\subsection{Enumerated List}
-\\begin{enumerate}
-\\item First item
-\\item Second item
-\\item Third item
-\\end{enumerate}",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "loc": {
-        "lines": {
-          "from": 27,
-          "to": 61,
-        },
-      },
-    },
-    "pageContent": "\\section{Lists}
-\\subsection{Itemized List}
-\\begin{itemize}
-\\item First item
-\\item Second item
-\\item Third item
-\\end{itemize}
-
-
-\\subsection{Enumerated List}
-\\begin{enumerate}
-\\item First item
-\\item Second item
-\\item Third item
-\\end{enumerate}
-
-
-\\section{Mathematical Equations}
-Here are some sample mathematical equations:
-
-
-\\subsection{Inline Equation}
-This is an inline equation: \\( E = mc^2 \\).
-
-
-\\subsection{Displayed Equations}
-\\begin{equation}
-a^2 + b^2 = c^2
-\\end{equation}
-
-
-\\begin{align}
-x &= y + z \\\\
-y &= mx + b
-\\end{align}",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "loc": {
-        "lines": {
-          "from": 44,
-          "to": 93,
-        },
-      },
-    },
-    "pageContent": "\\section{Mathematical Equations}
-Here are some sample mathematical equations:
-
-
-\\subsection{Inline Equation}
-This is an inline equation: \\( E = mc^2 \\).
-
-
-\\subsection{Displayed Equations}
-\\begin{equation}
-a^2 + b^2 = c^2
-\\end{equation}
-
-
-\\begin{align}
-x &= y + z \\\\
-y &= mx + b
-\\end{align}
-
-
-\\section{Tables}
-Here is a sample table:
-
-
-\\begin{table}[h!]
-\\centering
-\\begin{tabular}{|c|c|c|}
-\\hline
-Header 1 & Header 2 & Header 3 \\\\
-\\hline
-Data 1 & Data 2 & Data 3 \\\\
-Data 4 & Data 5 & Data 6 \\\\
-Data 7 & Data 8 & Data 9 \\\\
-\\hline
-\\end{tabular}
-\\caption{Sample Table}
-\\label{table:1}
-\\end{table}
-
-
-\\section{Figures}
-Here is a sample figure:
-
-
-\\begin{figure}[h!]
-\\centering
-\\includegraphics[width=0.5\\textwidth]{example-image}
-\\caption{Sample Figure}
-\\label{fig:1}
-\\end{figure}",
-  },
-  Document {
-    "id": undefined,
-    "metadata": {
-      "loc": {
-        "lines": {
-          "from": 84,
-          "to": 112,
-        },
-      },
-    },
-    "pageContent": "\\section{Figures}
-Here is a sample figure:
-
-
-\\begin{figure}[h!]
-\\centering
-\\includegraphics[width=0.5\\textwidth]{example-image}
-\\caption{Sample Figure}
-\\label{fig:1}
-\\end{figure}
-
-
-\\section{Sections and Subsections}
-This is an example of a section with subsections.
-
-
-\\subsection{Subsection 1}
-Content of subsection 1.
-
-
-\\subsection{Subsection 2}
-Content of subsection 2.
-
-
-\\section{References}
-Here is a reference to the table \\ref{table:1} and the figure \\ref{fig:1}.
-
-
-\\end{document}",
-  },
-]
-`;
@@ -1,9 +0,0 @@
-import { LatexTextSplitter } from 'langchain/text_splitter';
-
-import { loaderConfig } from '../config';
-
-export const LatexLoader = async (text: string) => {
-  const splitter = new LatexTextSplitter(loaderConfig);
-
-  return await splitter.createDocuments([text]);
-};
@@ -1,9 +0,0 @@
-import { MarkdownTextSplitter } from 'langchain/text_splitter';
-
-import { loaderConfig } from '../config';
-
-export const MarkdownLoader = async (text: string) => {
-  const splitter = new MarkdownTextSplitter(loaderConfig);
-
-  return await splitter.createDocuments([text]);
-};
@@ -1,7 +0,0 @@
-import { PDFLoader } from '@langchain/community/document_loaders/fs/pdf';
-
-export const PdfLoader = async (fileBlob: Blob) => {
-  const loader = new PDFLoader(fileBlob, { splitPages: true });
-
-  return await loader.load();
-};
@@ -1,7 +0,0 @@
-import { PPTXLoader as Loader } from '@langchain/community/document_loaders/fs/pptx';
-
-export const PPTXLoader = async (fileBlob: Blob | string) => {
-  const loader = new Loader(fileBlob);
-
-  return await loader.load();
-};
@@ -1,9 +0,0 @@
-import { RecursiveCharacterTextSplitter } from 'langchain/text_splitter';
-
-import { loaderConfig } from '../config';
-
-export const TextLoader = async (text: string) => {
-  const splitter = new RecursiveCharacterTextSplitter(loaderConfig);
-
-  return await splitter.createDocuments([text]);
-};
@@ -1,10 +0,0 @@
-export type LangChainLoaderType =
-  | 'code'
-  | 'ppt'
-  | 'pdf'
-  | 'markdown'
-  | 'doc'
-  | 'text'
-  | 'latex'
-  | 'csv'
-  | 'epub';
@@ -104,12 +104,22 @@ export default {
    'The request returned empty. Please check if the API proxy address does not end with `/v1`.',
  'response.CreateMessageError':
    'Sorry, the message could not be sent successfully. Please copy the content and try sending it again. This message will not be retained after refreshing the page.',
+  'exceededContext.compact': 'Compact Context',
+  'exceededContext.desc':
+    'The conversation has exceeded the context window limit. You can compact the context to compress history and continue chatting.',
+  'exceededContext.title': 'Context Window Exceeded',
+
+  'unknownError.copyTraceId': 'Trace ID Copied',
+  'unknownError.desc': 'An unexpected error occurred. You can retry or report on',
+  'unknownError.retry': 'Retry',
+  'unknownError.title': 'Oops, the request took a nap',
+
  'response.ExceededContextWindow':
    'The current request content exceeds the length that the model can handle. Please reduce the amount of content and try again.',
  'response.ExceededContextWindowCloud':
    'The conversation is too long to process. Please edit your last message to reduce input or delete some messages and try again.',
  'response.QuotaLimitReachedCloud':
-    'The model service is currently under heavy load. Please try again later.',
+    'The model service is currently under heavy load. Please try again later or switch to another model.',
  'response.FreePlanLimit':
    'You are currently a free user and cannot use this feature. Please upgrade to a paid plan to continue using it.',
  'response.InsufficientBudgetForModel':
@@ -25,6 +25,7 @@ import metadata from './metadata';
 import migration from './migration';
 import modelProvider from './modelProvider';
 import models from './models';
+import notification from './notification';
 import oauth from './oauth';
 import onboarding from './onboarding';
 import plugin from './plugin';
@@ -72,6 +73,7 @@ const resources = {
  migration,
  modelProvider,
  models,
+  notification,
  oauth,
  onboarding,
  plugin,
@@ -0,0 +1,12 @@
+export default {
+  'image_generation_completed': 'Image "{{prompt}}" generated successfully',
+  'image_generation_completed_title': 'Image Generated',
+  'inbox.archiveAll': 'Archive all',
+  'inbox.empty': 'No notifications yet',
+  'inbox.emptyUnread': 'No unread notifications',
+  'inbox.filterUnread': 'Show unread only',
+  'inbox.markAllRead': 'Mark all as read',
+  'inbox.title': 'Notifications',
+  'video_generation_completed': 'Video "{{prompt}}" generated successfully',
+  'video_generation_completed_title': 'Video Generated',
+} as const;
@@ -450,6 +450,12 @@ export default {
  'memory.enabled.title': 'Enable Memory',
  'memory.title': 'Memory Settings',
  'message.success': 'Update successful',
+  'notification.enabled': 'Enabled',
+  'notification.email.desc': 'Receive email notifications when important events occur',
+  'notification.email.title': 'Email Notifications',
+  'notification.inbox.desc': 'Show notifications in the in-app inbox',
+  'notification.inbox.title': 'Inbox Notifications',
+  'notification.title': 'Notification Channels',
  'myAgents.actions.cancel': 'Cancel',
  'myAgents.actions.confirmDeprecate': 'Confirm Deprecate',
  'myAgents.actions.deprecate': 'Deprecate Permanently',
@@ -927,6 +933,7 @@ When I am ___, I need ___
  'tab.manualFill': 'Manually Fill In',
  'tab.manualFill.desc': 'Configure a custom MCP skill manually',
  'tab.memory': 'Memory',
+  'tab.notification': 'Notifications',
  'tab.profile': 'My Account',
  'tab.provider': 'Provider',
  'tab.proxy': 'Proxy',
@@ -22,7 +22,7 @@ export const LoadingState = memo<LoadingStateProps>(
    return (
      <Block
        align={'center'}
-        className={styles.placeholderContainer}
+        className={`${styles.placeholderContainer} ${styles.placeholderContainerLoading}`}
        justify={'center'}
        variant={'filled'}
        style={{
@@ -30,7 +30,7 @@ export const LoadingState = memo<LoadingStateProps>(
          maxWidth: getThumbnailMaxWidth(generation, generationBatch),
        }}
      >
-        <div className={styles.placeholderContainer} />
+        <div className={`${styles.placeholderContainer} ${styles.placeholderContainerLoading}`} />
        <Center gap={8} style={{ zIndex: 2 }}>
          <NeuralNetworkLoading size={48} />
          <ElapsedTime generationId={generation.id} isActive={isGenerating} />
@@ -43,6 +43,12 @@ export const styles = createStaticStyles(({ css, cssVar }) => ({
    overflow: hidden;
    width: 100%;

+    &:hover .generation-actions {
+      opacity: 1;
+    }
+  `,
+
+  placeholderContainerLoading: css`
    &::before {
      content: '';

@@ -54,10 +60,6 @@ export const styles = createStaticStyles(({ css, cssVar }) => ({

      animation: ${shimmer} 2s linear infinite;
    }
-
-    &:hover .generation-actions {
-      opacity: 1;
-    }
  `,

  spinIcon: css`
@@ -0,0 +1,53 @@
+'use client';
+
+import { ActionIcon } from '@lobehub/ui';
+import { Badge } from 'antd';
+import { BellIcon } from 'lucide-react';
+import { memo, useCallback, useState } from 'react';
+import { useTranslation } from 'react-i18next';
+
+import { DESKTOP_HEADER_ICON_SIZE } from '@/const/layoutTokens';
+import { useClientDataSWR } from '@/libs/swr';
+import { notificationService } from '@/services/notification';
+import { serverConfigSelectors, useServerConfigStore } from '@/store/serverConfig';
+
+import InboxDrawer from './InboxDrawer';
+import { UNREAD_COUNT_KEY } from './InboxDrawer/constants';
+
+const InboxButton = memo(() => {
+  const { t } = useTranslation('notification');
+  const [open, setOpen] = useState(false);
+  const enableBusinessFeatures = useServerConfigStore(serverConfigSelectors.enableBusinessFeatures);
+
+  const { data: unreadCount = 0 } = useClientDataSWR<number>(
+    enableBusinessFeatures ? UNREAD_COUNT_KEY : null,
+    () => notificationService.getUnreadCount(),
+    { refreshInterval: 10_000 },
+  );
+
+  const handleToggle = useCallback(() => {
+    setOpen((prev) => !prev);
+  }, []);
+
+  const handleClose = useCallback(() => {
+    setOpen(false);
+  }, []);
+
+  if (!enableBusinessFeatures) return null;
+
+  return (
+    <>
+      <Badge count={unreadCount} offset={[-4, 4]} size="small">
+        <ActionIcon
+          icon={BellIcon}
+          size={DESKTOP_HEADER_ICON_SIZE}
+          title={t('inbox.title')}
+          onClick={handleToggle}
+        />
+      </Badge>
+      <InboxDrawer open={open} onClose={handleClose} />
+    </>
+  );
+});
+
+export default InboxButton;
@@ -0,0 +1,120 @@
+'use client';
+
+import { Flexbox, Icon, Text } from '@lobehub/ui';
+import { cssVar } from 'antd-style';
+import { BellOffIcon } from 'lucide-react';
+import { memo, useCallback, useEffect, useRef } from 'react';
+import { useTranslation } from 'react-i18next';
+import useSWRInfinite from 'swr/infinite';
+import { VList, type VListHandle } from 'virtua';
+
+import SkeletonList from '@/features/NavPanel/components/SkeletonList';
+import { notificationService } from '@/services/notification';
+
+import { FETCH_KEY } from './constants';
+import NotificationItem from './NotificationItem';
+
+const PAGE_SIZE = 20;
+
+interface ContentProps {
+  onArchive: (id: string) => void;
+  onMarkAsRead: (id: string) => void;
+  open: boolean;
+  unreadOnly?: boolean;
+}
+
+const Content = memo<ContentProps>(({ open, unreadOnly, onMarkAsRead, onArchive }) => {
+  const { t } = useTranslation('notification');
+  const virtuaRef = useRef<VListHandle>(null);
+
+  const getKey = useCallback(
+    (pageIndex: number, previousPageData: any[] | null) => {
+      if (!open) return null;
+      if (previousPageData && previousPageData.length < PAGE_SIZE) return null;
+
+      if (pageIndex === 0) return [FETCH_KEY, undefined, unreadOnly] as const;
+
+      const lastItem = previousPageData?.at(-1);
+      return [FETCH_KEY, lastItem?.id, unreadOnly] as const;
+    },
+    [open, unreadOnly],
+  );
+
+  const {
+    data: pages,
+    isLoading,
+    isValidating,
+    setSize,
+  } = useSWRInfinite(getKey, async ([, cursor, filterUnread]) => {
+    return notificationService.list({
+      cursor: cursor as string | undefined,
+      limit: PAGE_SIZE,
+      unreadOnly: filterUnread,
+    });
+  });
+
+  // Reset scroll position and pagination when filter changes
+  useEffect(() => {
+    setSize(1);
+    virtuaRef.current?.scrollTo(0);
+  }, [unreadOnly, setSize]);
+
+  const notifications = pages?.flat() ?? [];
+  const hasMore = pages ? pages.at(-1)?.length === PAGE_SIZE : false;
+
+  const handleScroll = useCallback(() => {
+    const ref = virtuaRef.current;
+    if (!ref || !hasMore || isValidating) return;
+
+    const bottomVisibleIndex = ref.findItemIndex(ref.scrollOffset + ref.viewportSize);
+    if (bottomVisibleIndex + 5 > notifications.length) {
+      setSize((prev) => prev + 1);
+    }
+  }, [hasMore, isValidating, notifications.length, setSize]);
+
+  if (isLoading) {
+    return (
+      <Flexbox gap={1} paddingBlock={1} paddingInline={4}>
+        <SkeletonList rows={5} />
+      </Flexbox>
+    );
+  }
+
+  if (notifications.length === 0) {
+    return (
+      <Flexbox align="center" gap={12} justify="center" paddingBlock={48}>
+        <Icon color={cssVar.colorTextQuaternary} icon={BellOffIcon} size={40} />
+        <Text type="secondary">{t(unreadOnly ? 'inbox.emptyUnread' : 'inbox.empty')}</Text>
+      </Flexbox>
+    );
+  }
+
+  return (
+    <VList ref={virtuaRef} style={{ height: '100%' }} onScroll={handleScroll}>
+      {notifications.map((item) => (
+        <Flexbox key={item.id} padding="4px 8px">
+          <NotificationItem
+            actionUrl={item.actionUrl}
+            content={item.content}
+            createdAt={item.createdAt}
+            id={item.id}
+            isRead={item.isRead}
+            title={item.title}
+            type={item.type}
+            onArchive={onArchive}
+            onMarkAsRead={onMarkAsRead}
+          />
+        </Flexbox>
+      ))}
+      {isValidating && (
+        <Flexbox padding="4px 8px">
+          <SkeletonList rows={2} />
+        </Flexbox>
+      )}
+    </VList>
+  );
+});
+
+Content.displayName = 'InboxDrawerContent';
+
+export default Content;
@@ -0,0 +1,122 @@
+'use client';
+
+import { ActionIcon, Block, Flexbox, Icon, Text } from '@lobehub/ui';
+import { createStaticStyles, cssVar } from 'antd-style';
+import dayjs from 'dayjs';
+import { ArchiveIcon, BellIcon, ImageIcon, VideoIcon } from 'lucide-react';
+import { memo, useCallback } from 'react';
+import { useNavigate } from 'react-router-dom';
+
+const ACTION_CLASS_NAME = 'notification-item-actions';
+
+const styles = createStaticStyles(({ css }) => ({
+  container: css`
+    cursor: pointer;
+    user-select: none;
+
+    .${ACTION_CLASS_NAME} {
+      opacity: 0;
+      transition: opacity 0.2s ${cssVar.motionEaseOut};
+    }
+
+    &:hover {
+      .${ACTION_CLASS_NAME} {
+        opacity: 1;
+      }
+    }
+  `,
+  unreadDot: css`
+    flex-shrink: 0;
+
+    width: 8px;
+    height: 8px;
+    border-radius: 50%;
+
+    background: ${cssVar.colorPrimary};
+  `,
+}));
+
+const TYPE_ICON_MAP: Record<string, typeof BellIcon> = {
+  image_generation_completed: ImageIcon,
+  video_generation_completed: VideoIcon,
+};
+
+interface NotificationItemProps {
+  actionUrl?: string | null;
+  content: string;
+  createdAt: Date | string;
+  id: string;
+  isRead: boolean;
+  onArchive: (id: string) => void;
+  onMarkAsRead: (id: string) => void;
+  title: string;
+  type: string;
+}
+
+const NotificationItem = memo<NotificationItemProps>(
+  ({ id, type, title, content, createdAt, isRead, actionUrl, onMarkAsRead, onArchive }) => {
+    const navigate = useNavigate();
+    const TypeIcon = TYPE_ICON_MAP[type] || BellIcon;
+
+    const handleClick = useCallback(() => {
+      if (!isRead) onMarkAsRead(id);
+      if (actionUrl) navigate(actionUrl);
+    }, [id, isRead, actionUrl, onMarkAsRead, navigate]);
+
+    const handleArchive = useCallback(
+      (e: React.MouseEvent) => {
+        e.stopPropagation();
+        onArchive(id);
+      },
+      [id, onArchive],
+    );
+
+    return (
+      <Block
+        clickable
+        className={styles.container}
+        gap={4}
+        paddingBlock={8}
+        paddingInline={12}
+        variant="borderless"
+        onClick={handleClick}
+      >
+        <Flexbox horizontal align="flex-start" gap={8}>
+          <Icon
+            color={cssVar.colorTextDescription}
+            icon={TypeIcon}
+            size={18}
+            style={{ flexShrink: 0, marginTop: 2 }}
+          />
+          <Flexbox flex={1} gap={4} style={{ overflow: 'hidden' }}>
+            <Flexbox horizontal align="center" gap={4} justify="space-between">
+              <Flexbox horizontal align="center" flex={1} gap={6} style={{ overflow: 'hidden' }}>
+                {!isRead && <span className={styles.unreadDot} />}
+                <Text ellipsis style={{ fontWeight: isRead ? 400 : 600 }}>
+                  {title}
+                </Text>
+              </Flexbox>
+              <Flexbox horizontal align="center" gap={2} style={{ flexShrink: 0 }}>
+                <span className={ACTION_CLASS_NAME}>
+                  <ActionIcon
+                    icon={ArchiveIcon}
+                    size={{ blockSize: 24, size: 14 }}
+                    onClick={handleArchive}
+                  />
+                </span>
+                <Text fontSize={12} style={{ flexShrink: 0 }} type="secondary">
+                  {dayjs(createdAt).fromNow()}
+                </Text>
+              </Flexbox>
+            </Flexbox>
+            <Text ellipsis={{ rows: 3 }} fontSize={12} type="secondary">
+              {content}
+            </Text>
+          </Flexbox>
+        </Flexbox>
+      </Block>
+    );
+  },
+);
+
+export default NotificationItem;
@@ -0,0 +1,2 @@
+export const UNREAD_COUNT_KEY = 'inbox-unread-count';
+export const FETCH_KEY = 'inbox-notifications';
@@ -0,0 +1,111 @@
+'use client';
+
+import { ActionIcon, Flexbox } from '@lobehub/ui';
+import { ArchiveIcon, CheckCheckIcon, ListFilterIcon } from 'lucide-react';
+import { memo, useCallback, useState } from 'react';
+import { useTranslation } from 'react-i18next';
+
+import { DESKTOP_HEADER_ICON_SIZE } from '@/const/layoutTokens';
+import SkeletonList from '@/features/NavPanel/components/SkeletonList';
+import SideBarDrawer from '@/features/NavPanel/SideBarDrawer';
+import dynamic from '@/libs/next/dynamic';
+import { mutate } from '@/libs/swr';
+import { notificationService } from '@/services/notification';
+
+import { FETCH_KEY, UNREAD_COUNT_KEY } from './constants';
+
+const Content = dynamic(() => import('./Content'), {
+  loading: () => (
+    <Flexbox gap={1} paddingBlock={1} paddingInline={4}>
+      <SkeletonList rows={3} />
+    </Flexbox>
+  ),
+  ssr: false,
+});
+
+interface InboxDrawerProps {
+  onClose: () => void;
+  open: boolean;
+}
+
+const InboxDrawer = memo<InboxDrawerProps>(({ open, onClose }) => {
+  const { t } = useTranslation('notification');
+  const [unreadOnly, setUnreadOnly] = useState(false);
+
+  const refreshList = useCallback(() => {
+    mutate((key: unknown) => Array.isArray(key) && key[0] === FETCH_KEY);
+    mutate(UNREAD_COUNT_KEY);
+  }, []);
+
+  const handleMarkAsRead = useCallback(
+    async (id: string) => {
+      await notificationService.markAsRead([id]);
+      refreshList();
+    },
+    [refreshList],
+  );
+
+  const handleArchive = useCallback(
+    async (id: string) => {
+      await notificationService.archive(id);
+      refreshList();
+    },
+    [refreshList],
+  );
+
+  const handleMarkAllAsRead = useCallback(async () => {
+    await notificationService.markAllAsRead();
+    refreshList();
+  }, [refreshList]);
+
+  const handleArchiveAll = useCallback(async () => {
+    await notificationService.archiveAll();
+    refreshList();
+  }, [refreshList]);
+
+  const handleToggleFilter = useCallback(() => {
+    setUnreadOnly((prev) => !prev);
+  }, []);
+
+  return (
+    <SideBarDrawer
+      open={open}
+      title={t('inbox.title')}
+      action={
+        <>
+          <ActionIcon
+            icon={ArchiveIcon}
+            size={DESKTOP_HEADER_ICON_SIZE}
+            title={t('inbox.archiveAll')}
+            onClick={handleArchiveAll}
+          />
+          <ActionIcon
+            icon={CheckCheckIcon}
+            size={DESKTOP_HEADER_ICON_SIZE}
+            title={t('inbox.markAllRead')}
+            onClick={handleMarkAllAsRead}
+          />
+          <ActionIcon
+            active={unreadOnly}
+            icon={ListFilterIcon}
+            size={DESKTOP_HEADER_ICON_SIZE}
+            title={t('inbox.filterUnread')}
+            onClick={handleToggleFilter}
+          />
+        </>
+      }
+      onClose={onClose}
+    >
+      <Content
+        open={open}
+        unreadOnly={unreadOnly}
+        onArchive={handleArchive}
+        onMarkAsRead={handleMarkAsRead}
+      />
+    </SideBarDrawer>
+  );
+});
+
+InboxDrawer.displayName = 'InboxDrawer';
+
+export default InboxDrawer;
@@ -5,13 +5,23 @@ import { memo } from 'react';
 import SideBarHeaderLayout from '@/features/NavPanel/SideBarHeaderLayout';

 import AddButton from './components/AddButton';
+import InboxButton from './components/InboxButton';
 import Nav from './components/Nav';
 import User from './components/User';

 const Header = memo(() => {
  return (
    <>
-      <SideBarHeaderLayout left={<User />} right={<AddButton />} showBack={false} />
+      <SideBarHeaderLayout
+        left={<User />}
+        showBack={false}
+        right={
+          <>
+            <InboxButton />
+            <AddButton />
+          </>
+        }
+      />
      <Nav />
    </>
  );
@@ -1,5 +1,6 @@
 import Billing from '@/business/client/BusinessSettingPages/Billing';
 import Credits from '@/business/client/BusinessSettingPages/Credits';
+import Notification from '@/business/client/BusinessSettingPages/Notification';
 import Plans from '@/business/client/BusinessSettingPages/Plans';
 import Referral from '@/business/client/BusinessSettingPages/Referral';
 import Usage from '@/business/client/BusinessSettingPages/Usage';
@@ -28,6 +29,7 @@ export const componentMap = {
  [SettingsTabs.Provider]: Provider,
  [SettingsTabs.ServiceModel]: ServiceModel,
  [SettingsTabs.Memory]: Memory,
+  [SettingsTabs.Notification]: Notification,
  [SettingsTabs.About]: About,
  [SettingsTabs.Hotkey]: Hotkey,
  [SettingsTabs.Proxy]: Proxy,
@@ -22,6 +22,12 @@ export const componentMap = {
  [SettingsTabs.Memory]: dynamic(() => import('../memory'), {
    loading: loading('Settings > Memory'),
  }),
+  [SettingsTabs.Notification]: dynamic(
+    () => import('@/business/client/BusinessSettingPages/Notification'),
+    {
+      loading: loading('Settings > Notification'),
+    },
+  ),
  [SettingsTabs.About]: dynamic(() => import('../about'), {
    loading: loading('Settings > About'),
  }),
@@ -2,6 +2,7 @@ import { isDesktop } from '@lobechat/const';
 import { Avatar } from '@lobehub/ui';
 import { SkillsIcon } from '@lobehub/ui/icons';
 import {
+  // BellIcon,
  Brain,
  BrainCircuit,
  ChartColumnBigIcon,
@@ -100,6 +101,12 @@ export const useCategory = () => {
        key: SettingsTabs.Hotkey,
        label: t('tab.hotkey'),
      },
+      // TODO: temporarily disabled until notification UI is polished
+      // enableBusinessFeatures && {
+      //   icon: BellIcon,
+      //   key: SettingsTabs.Notification,
+      //   label: t('tab.notification'),
+      // },
    ].filter(Boolean) as CategoryItem[];

    groups.push({
@@ -127,7 +134,7 @@ export const useCategory = () => {

    // Agent group
    const agentItems: CategoryItem[] = [
-      {
+      (!enableBusinessFeatures || isDevMode) && {
        icon: Brain,
        key: SettingsTabs.Provider,
        label: t('tab.provider'),
@@ -1,12 +1,12 @@
 import { beforeEach, describe, expect, it, vi } from 'vitest';

 import { knowledgeEnv } from '@/envs/knowledge';
-import { ChunkingLoader } from '@/libs/langchain';
+import { ChunkingLoader } from '@/libs/document-loaders';

 import { ContentChunk } from './index';

 // Mock the dependencies
-vi.mock('@/libs/langchain');
+vi.mock('@/libs/document-loaders');
 vi.mock('@/envs/knowledge', () => ({
  knowledgeEnv: {
    FILE_TYPE_CHUNKING_RULES: '',
@@ -70,7 +70,7 @@ describe('ContentChunk', () => {
        index: 0,
        metadata: { source: 'test' },
        text: 'Test content chunk 1',
-        type: 'LangChainElement',
+        type: 'DocumentChunk',
      });
      expect(result.unstructuredChunks).toBeUndefined();
    });
@@ -143,13 +143,13 @@ describe('ContentChunk', () => {
          loc: { lines: { from: 1, to: 10 } },
        },
        text: 'First paragraph content',
-        type: 'LangChainElement',
+        type: 'DocumentChunk',
      });
      expect(result.chunks[1]).toMatchObject({
        id: 'chunk-2',
        index: 1,
        text: 'Second paragraph content',
-        type: 'LangChainElement',
+        type: 'DocumentChunk',
      });
    });

@@ -242,7 +242,7 @@ describe('ContentChunk', () => {
        index: 0,
        metadata: {},
        text: 'Content with no metadata',
-        type: 'LangChainElement',
+        type: 'DocumentChunk',
      });
    });
  });
@@ -1,6 +1,6 @@
 import { type NewChunkItem, type NewUnstructuredChunkItem } from '@/database/schemas';
 import { knowledgeEnv } from '@/envs/knowledge';
-import { ChunkingLoader } from '@/libs/langchain';
+import { ChunkingLoader } from '@/libs/document-loaders';

 import { type ChunkingService } from './rules';
 import { ChunkingRuleParser } from './rules';
@@ -18,11 +18,11 @@ interface ChunkResult {
 }

 export class ContentChunk {
-  private langchainClient: ChunkingLoader;
+  private chunkingClient: ChunkingLoader;
  private chunkingRules: Record<string, ChunkingService[]>;

  constructor() {
-    this.langchainClient = new ChunkingLoader();
+    this.chunkingClient = new ChunkingLoader();
    this.chunkingRules = ChunkingRuleParser.parse(knowledgeEnv.FILE_TYPE_CHUNKING_RULES || '');
  }

@@ -43,7 +43,7 @@ export class ContentChunk {
          }

          default: {
-            return await this.chunkByLangChain(params.filename, params.content);
+            return await this.chunkByDefault(params.filename, params.content);
          }
        }
      } catch (error) {
@@ -54,26 +54,23 @@ export class ContentChunk {
      }
    }

-    // Fallback to langchain if no service succeeded
-    return await this.chunkByLangChain(params.filename, params.content);
+    // Fallback to default chunking if no service succeeded
+    return await this.chunkByDefault(params.filename, params.content);
  }

  private canUseUnstructured(): boolean {
    return !!(knowledgeEnv.UNSTRUCTURED_API_KEY && knowledgeEnv.UNSTRUCTURED_SERVER_URL);
  }

-  private chunkByLangChain = async (
-    filename: string,
-    content: Uint8Array,
-  ): Promise<ChunkResult> => {
-    const res = await this.langchainClient.partitionContent(filename, content);
+  private chunkByDefault = async (filename: string, content: Uint8Array): Promise<ChunkResult> => {
+    const res = await this.chunkingClient.partitionContent(filename, content);

    const documents = res.map((item, index) => ({
      id: item.id,
      index,
      metadata: item.metadata,
      text: item.pageContent,
-      type: 'LangChainElement',
+      type: 'DocumentChunk',
    }));

    return { chunks: documents };
@@ -8,6 +8,7 @@ import { type RuntimeImageGenParams } from 'model-bank';
 import { z } from 'zod';

 import { chargeAfterGenerate } from '@/business/server/image-generation/chargeAfterGenerate';
+import { notifyImageCompleted } from '@/business/server/image-generation/notifyImageCompleted';
 import { createImageBusinessMiddleware } from '@/business/server/trpc-middlewares/async';
 import { AsyncTaskModel } from '@/database/models/asyncTask';
 import { FileModel } from '@/database/models/file';
@@ -155,6 +156,22 @@ const categorizeError = (
    };
  }

+  // Content moderation / policy violation — return a clean, generic message
+  const errorMsg: string = error.message || error.error?.message || '';
+  const errorCode: string = error.code || error.error?.code || '';
+  if (
+    errorCode === 'InputTextSensitiveContentDetected' ||
+    errorCode === 'content_policy_violation' ||
+    errorMsg.toLowerCase().includes('content policy') ||
+    errorMsg.toLowerCase().includes('sensitive information')
+  ) {
+    return {
+      errorMessage:
+        'The request content may violate content policy. Please modify your prompt and try again.',
+      errorType: AsyncTaskErrorType.ServerError,
+    };
+  }
+
  if (error instanceof AsyncTaskError) {
    return {
      errorMessage: typeof error.body === 'string' ? error.body : error.body.detail,
@@ -342,6 +359,15 @@ export const imageRouter = router({
            status: AsyncTaskStatus.Success,
          });

+          notifyImageCompleted({
+            duration,
+            generationBatchId,
+            model,
+            prompt: params.prompt,
+            topicId: generationTopicId,
+            userId: ctx.userId,
+          }).catch((err) => console.error('[image-async] notification failed:', err));
+
          if (ENABLE_BUSINESS_FEATURES) {
            await chargeAfterGenerate({
              metrics: { latency: duration },
@@ -41,6 +41,7 @@ import { knowledgeBaseRouter } from './knowledgeBase';
 import { marketRouter } from './market';
 import { messageRouter } from './message';
 import { notebookRouter } from './notebook';
+import { notificationRouter } from './notification';
 import { oauthDeviceFlowRouter } from './oauthDeviceFlow';
 import { pluginRouter } from './plugin';
 import { ragEvalRouter } from './ragEval';
@@ -94,6 +95,7 @@ export const lambdaRouter = router({
  market: marketRouter,
  message: messageRouter,
  notebook: notebookRouter,
+  notification: notificationRouter,
  oauthDeviceFlow: oauthDeviceFlowRouter,
  plugin: pluginRouter,
  ragEval: ragEvalRouter,
@@ -0,0 +1,54 @@
+import { z } from 'zod';
+
+import { NotificationModel } from '@/database/models/notification';
+import { authedProcedure, router } from '@/libs/trpc/lambda';
+import { serverDatabase } from '@/libs/trpc/lambda/middleware';
+
+const notificationProcedure = authedProcedure.use(serverDatabase).use(async (opts) => {
+  const { ctx } = opts;
+
+  return opts.next({
+    ctx: { notificationModel: new NotificationModel(ctx.serverDB, ctx.userId) },
+  });
+});
+
+export const notificationRouter = router({
+  archive: notificationProcedure
+    .input(z.object({ id: z.string() }))
+    .mutation(async ({ ctx, input }) => {
+      return ctx.notificationModel.archive(input.id);
+    }),
+
+  archiveAll: notificationProcedure.mutation(async ({ ctx }) => {
+    return ctx.notificationModel.archiveAll();
+  }),
+
+  list: notificationProcedure
+    .input(
+      z.object({
+        category: z.string().optional(),
+        cursor: z.string().optional(),
+        limit: z.number().min(1).max(50).default(20),
+        unreadOnly: z.boolean().optional(),
+      }),
+    )
+    .query(async ({ ctx, input }) => {
+      return ctx.notificationModel.list(input);
+    }),
+
+  markAllAsRead: notificationProcedure.mutation(async ({ ctx }) => {
+    return ctx.notificationModel.markAllAsRead();
+  }),
+
+  markAsRead: notificationProcedure
+    .input(z.object({ ids: z.array(z.string()).min(1) }))
+    .mutation(async ({ ctx, input }) => {
+      return ctx.notificationModel.markAsRead(input.ids);
+    }),
+
+  unreadCount: notificationProcedure.query(async ({ ctx }) => {
+    return ctx.notificationModel.getUnreadCount();
+  }),
+});
+
+export type NotificationRouter = typeof notificationRouter;
@@ -0,0 +1,36 @@
+import { lambdaClient } from '@/libs/trpc/client';
+
+class NotificationService {
+  list = (
+    params: {
+      category?: string;
+      cursor?: string;
+      limit?: number;
+      unreadOnly?: boolean;
+    } = {},
+  ) => {
+    return lambdaClient.notification.list.query(params);
+  };
+
+  getUnreadCount = (): Promise<number> => {
+    return lambdaClient.notification.unreadCount.query();
+  };
+
+  markAsRead = (ids: string[]) => {
+    return lambdaClient.notification.markAsRead.mutate({ ids });
+  };
+
+  markAllAsRead = () => {
+    return lambdaClient.notification.markAllAsRead.mutate();
+  };
+
+  archive = (id: string) => {
+    return lambdaClient.notification.archive.mutate({ id });
+  };
+
+  archiveAll = () => {
+    return lambdaClient.notification.archiveAll.mutate();
+  };
+}
+
+export const notificationService = new NotificationService();
@@ -149,7 +149,7 @@ export class ConversationLifecycleActionImpl {
        compressContext.topicId &&
        !hasRunningCompressionOperation(Object.values(this.#get().operations), compressContext)
      ) {
-        await this.#executeCompression(compressContext, '');
+        await this.executeCompression(compressContext, '');
      }
      return;
    }
@@ -596,7 +596,7 @@ export class ConversationLifecycleActionImpl {
   * Execute context compression for /compact command.
   * Reuses the same service methods as the agent runtime's compress_context executor.
   */
-  #executeCompression = async (
+  executeCompression = async (
    context: Record<string, any>,
    parentOperationId: string,
  ): Promise<void> => {
@@ -170,8 +170,7 @@ export class FileActionImpl {
      // image don't need to be chunked and embedding
      if (isChunkingUnsupported(file.type)) return;

-      const data = await ragService.parseFileContent(fileResult.id);
-      console.info('parseFileContent data:', data);
+      await ragService.parseFileContent(fileResult.id);
    });

    await Promise.all(pools);
@@ -56,6 +56,7 @@ export enum SettingsTabs {
  Image = 'image',
  LLM = 'llm',
  Memory = 'memory',
+  Notification = 'notification',
  // business
  Plans = 'plans',
  Profile = 'profile',
Author	SHA1	Message	Date
arvinxx	af50ae40fd	✅ test: add e2e regression test for topic switch page reload Add a regression test to verify that switching topics within the same agent does not trigger a full browser page reload. The test injects a window marker before switching and verifies it survives after the switch. Related: lobehub/lobehub#13309 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-03-26 21:31:19 +08:00
Arvin Xu	0724d8ca60	🐛 fix: prevent full page reload when switching topics during agent execution (#13309 ) Move `e.preventDefault()` before the `disabled \|\| loading` early return in NavItem's onClick handler. Previously, when a NavItem was in disabled or loading state, the early return skipped `preventDefault()`, allowing the underlying `<a>` tag's default navigation to trigger a full browser page load instead of SPA routing. Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-03-26 21:30:08 +08:00
YuTengjing	9f36fe95ac	✨ feat: add notification system (temporarily disabled) (#13301 )	2026-03-26 21:16:38 +08:00
Arvin Xu	3f148005e4	♻️ refactor: remove langchain dependency, use direct document loaders (#13304 ) * ♻️ refactor: remove langchain dependency, use direct document loaders Replace langchain and @langchain/community with self-implemented text splitters and direct usage of underlying libraries (pdf-parse, d3-dsv, mammoth, officeparser, epub2). This eliminates unnecessary dependency bloat and addresses CVE-2026-26019 in @langchain/community. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * 🐛 fix: add missing @types/html-to-text and @types/pdf-parse Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-03-26 21:13:55 +08:00
Arvin Xu	4e60d87514	🔒 refactor: remove deprecated SystemJS plugin renderer (#13305 ) 🔒 fix: remove SystemJS plugin renderer to eliminate arbitrary JS execution risk The old plugin render system (ui.mode === 'module') that used SystemJS to dynamically load and execute JS from untrusted URLs has been fully retired. Remove SystemJsRender and systemjs dependency entirely. Ref: GHSA-46v7-wvmj-6vf7 Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-03-26 20:41:06 +08:00
YuTengjing	d2a16d0714	✨ feat: improve error UI and error handling across chat and image generation (#13302 )	2026-03-26 20:09:06 +08:00