diff --git a/.env.example b/.env.example
index ce10296c..087fd4c0 100644
--- a/.env.example
+++ b/.env.example
@@ -68,6 +68,14 @@ WORKER_PUBLIC_BANK_SYNC_COPY_LIMIT=1000
 WORKER_PUBLIC_BANK_SYNC_ID=public-banks-1
 WORKER_PUBLIC_BANK_SYNC_CLAIM_STALE_SECONDS=900
 
+# Worker 配置：题库 PDF/Word 导出。生产建议使用真实对象存储 provider；local_dev 只用于本地开发。
+WORKER_EXPORT_BATCH_SIZE=5
+WORKER_EXPORT_ID=exports-1
+WORKER_EXPORT_BACKOFF_SECONDS=30,120,600,1800
+EXPORT_LOCAL_STORAGE_ROOT=.local-storage
+# Windows 本地可填 C:\Windows\Fonts\NotoSansSC-VF.ttf；Linux 生产建议安装 Noto Sans CJK 并配置路径。
+EXPORT_PDF_FONT_PATH=
+
 # 对象存储配置：API 和 assets worker 共用
 STORAGE_DEFAULT_PROVIDER=local_dev
 STORAGE_DEFAULT_BUCKET=tenant-assets
diff --git a/.gitignore b/.gitignore
index e5affc5e..cd418f08 100644
--- a/.gitignore
+++ b/.gitignore
@@ -38,6 +38,7 @@ scripts/satellite/sync-config.json
 # ✅ 临时/对比文件
 stash_comparison.txt
 stash_diff.txt
+.local-storage/
 
 # ✅ 备份文件
 *.backup
diff --git a/README.md b/README.md
index 6c19fffb..8ce62a84 100644
--- a/README.md
+++ b/README.md
@@ -17,9 +17,9 @@
 - 学生端能力：题库入口、分类树、题目集合、顺序/随机/模考 session 组卷快照、答题、错题本、收藏夹、背单词进度、个人中心、勋章、考试倒计时、签到积分、题目反馈、排行榜、分数线、题目视频、订单详情/状态轮询、优惠券领取/抵扣、权益、激活码预检查/兑换、资料下载。
 - 平台后台能力：租户管理、SaaS 套餐、订阅、账单、服务费收款、用量记录、公共题库授权。
 - 公共题库商业化能力：租户可采纳平台授权题库为本租户副本，并可手动或由 worker 自动同步平台新增/更新题目；同步会保护租户自改题目，返回冲突而不覆盖，后台可查询冲突明细。
-- 题库导出基础能力：租户内容编辑可按题目集合、内容入口或分类节点导出 JSON、`paper_json` 和打印 payload，后端强制租户隔离、答案/解析开关、复合题子题脱敏、导出 job 和审计。
+- 题库导出能力：租户内容编辑可按题目集合、内容入口或分类节点导出 JSON、`paper_json`、打印 payload、PDF 和 Word，后端强制租户隔离、答案/解析开关、复合题子题脱敏、导出 job 和审计；PDF/Word 由 exports worker 生成水印文件并发布到 `content_assets`。
 - 销售/代理/CRM 增长链路：邀请码、扫码/分享事件、首绑客资保护、销售统计、团队关系、CRM 配置和队列。
-- `apps/worker` 后台任务进程：CRM webhook 队列消费、generic/钉钉/飞书/企微机器人发送、签名、失败重试和日志；commerce worker 可补偿查询微信/支付宝支付和退款状态；assets worker 可复检托管资源元数据并自动下架异常资源；public-banks worker 可自动同步公共题库采纳副本。
+- `apps/worker` 后台任务进程：CRM webhook 队列消费、generic/钉钉/飞书/企微机器人发送、签名、失败重试和日志；commerce worker 可补偿查询微信/支付宝支付和退款状态；assets worker 可复检托管资源元数据并自动下架异常资源；imports worker 可执行大批量导入；public-banks worker 可自动同步公共题库采纳副本；exports worker 可渲染 PDF/Word 导出文件。
 - 销售/代理分佣结算基础闭环：租户默认比例、成员比例、激活码批次比例、订单/激活码归因、结算单生成、审核、线下打款状态和权限隔离。
 - 订单售后基础闭环：退款请求、审核、处理状态流、微信/支付宝发起退款、微信/支付宝退款查询确认、微信/支付宝退款通知 webhook、退款金额累计、部分/全额退款订单状态、全额退款权益撤销、退款事件和审计日志。
 - PocketBase schema/数据导入器雏形和导入后校验脚本。
@@ -31,7 +31,7 @@
 - 阿里云/腾讯云短信、微信小程序登录、微信网页登录、QQ 登录、手机号绑定/换绑、微信支付、支付宝主链路、微信/支付宝发起退款/查询确认/退款通知、支付/退款补偿 worker 已完成本地适配；完整资金流水对账和真实生产账号联调还没接完。
 - OSS/COS/Supabase Storage 上传下载签名 provider 已接入；上传后校验、PDF/图片预览和资源复检 worker 已完成，CDN 防盗链、杀毒扫描和视频动态水印还没完成。
 - Excel/CSV 导入解析已完成并复用 `content_import_jobs/items/issues` 管线；大批量异步导入 worker 基础已接入，支持 queued job 消费、重试和审计；导入后复检、模板下载和字段映射 API 已完成，前端 UI 待接。
-- 题库导出目前完成服务端结构化 payload；PDF/Word 二进制生成、导出水印、发布到资料下载和导出 worker 还没完成。
+- 题库导出已完成服务端结构化 payload 和 PDF/Word 二进制 worker；后续还要补更精细试卷模板、每日一练素材导出和导出操作台体验。
 - 勋章管理/手动发放已可联调；自动发放规则、积分活动联动、分佣真实打款、结算导出、发票/凭证、CRM 轮询/定向分配、富卡片模板、失败告警、销售转化看板、公共题库版本通知和冲突处理操作台还没完成。
 - `apps/taro` 已建立 Taro 4 React 跨端前端地基，包含 H5 学生端、租户后台、平台后台三套构建入口、租户解析、统一 API client 和 Supabase Auth client 初始化；学生端第一批页面已接入登录、首页、题库、练习、背单词、知识手册、分数线、资料和个人中心；租户后台第一批页面已接入工作台、数据看板、学生/班级、题库内容、营销中心和租户设置；平台后台已接入工作台、租户管理、账务中心、公共题库授权，以及创建租户、状态变更、订阅、账单、收款、用量和题库授权第一版写操作。
 - 根目录已清理为新 Supabase SaaS monorepo 编排层；旧 PocketBase/React 项目和旧构建产物仅保留在 `参考/` 目录作为迁移参考，不进入 Git 提交。
@@ -60,7 +60,7 @@
 ```text
 apps/api/                   Node.js 业务 API
 apps/taro/                  Taro 4 React 跨端前端，H5 三入口，后续扩展小程序
-apps/worker/                后台异步任务：CRM webhook、支付/退款补偿、资源复检、导入执行、公共题库同步等
+apps/worker/                后台异步任务：CRM webhook、支付/退款补偿、资源复检、导入执行、公共题库同步、题库导出渲染等
 packages/config/            共享配置
 packages/db/                PostgreSQL 连接池和查询封装
 packages/domain/            领域常量和共享类型
@@ -181,6 +181,12 @@ npm --workspace @tiku-saas/worker run imports:once
 npm --workspace @tiku-saas/worker run public-banks:once
 ```
 
+单次运行题库 PDF/Word 导出 worker：
+
+```bash
+npm --workspace @tiku-saas/worker run exports:once
+```
+
 默认本地数据库：
 
 ```text
@@ -226,6 +232,7 @@ npm run test:api
 npm run test:worker:crm
 npm run test:worker:commerce
 npm run test:worker:assets
+npm run test:worker:exports
 npm run test:worker:imports
 npm run test:worker:public-banks
 ```
@@ -300,7 +307,7 @@ API 身份上下文：
 - 资料、PDF、视频等资源必须先进入 `content_assets` 台账，再由 API 校验权限并下发签名 URL；生产环境应定时运行 assets worker 复检对象元数据，异常资源会被标记 failed 并退回 draft。
 - 题库入口和分类使用 `content_entries/content_nodes`；题目列表和练习规则使用 `question_collections/practice_blueprints`，前端不要再把旧树字段当成唯一业务结构。
 - 批量导入必须先写 `content_import_jobs/items/issues`，保留原始 payload、规范化 payload、逐行问题和审计记录。题目、单词、知识手册、分数线和视频 JSON/CSV/Excel 导入已走这套后台校验管线；大批量任务可提交 `executionMode=async`，由 imports worker 消费，前端只轮询 job 状态和展示 issues。
-- 题库导出必须由后端按权限生成，不允许前端直接读取数据库拼导出文件；不开启答案/解析时，顶层题目和复合题子题都必须脱敏。
+- 题库导出必须由后端按权限生成，不允许前端直接读取数据库拼导出文件；不开启答案/解析时，顶层题目和复合题子题都必须脱敏；PDF/Word 只通过 exports worker 写入 `content_assets` 后再签名下载/预览。
 - 支付 webhook 必须先设计幂等键和验签流程，再进入生产使用；生产环境还应定时运行 commerce worker 兜底供应商漏通知和处理中退款。
 
 ## 最近一次验证
@@ -316,6 +323,7 @@ npm run test:worker:imports
 npm run test:worker:crm
 npm run test:worker:commerce
 npm run test:worker:assets
+npm run test:worker:exports
 npm run test:worker:public-banks
 npm run check:taro
 npm run build:taro:h5:student
@@ -335,5 +343,5 @@ git diff --check
 1. 真实云端 Auth/JWKS 回归、RLS 深测和生产环境配置验收。
 2. 继续补 Taro 前端：学生端视频/反馈/模考报告/订单收银台，租户后台写入表单/导入操作台/公共题库同步/角色模板 UI，平台后台租户详情/审计/自动计费增强，小程序兼容验证。
 3. 对象存储 CDN 防盗链、杀毒扫描、视频动态水印和生命周期策略。
-4. 题库导出 PDF/Word worker、真实数据 dry-run、导入字段映射 UI 和复检结果操作台。
+4. 题库导出模板精排、每日一练运营工具、真实数据 dry-run、导入字段映射 UI 和复检结果操作台。
 5. 真实 OAuth/短信/支付生产账号联调、完整资金流水对账、公共题库版本通知/冲突处理操作台、积分活动深化，以及排行榜防刷/预聚合。
diff --git a/apps/api/src/features/tenant-content/exports.ts b/apps/api/src/features/tenant-content/exports.ts
index e1bd8893..afa4c3ed 100644
--- a/apps/api/src/features/tenant-content/exports.ts
+++ b/apps/api/src/features/tenant-content/exports.ts
@@ -6,7 +6,9 @@ import { query, transaction } from '../../core/db.js';
 import { requireTenantContentEditor, type TenantContentAuth } from './auth.js';
 import { boolValue, jsonObjectValue, nullableString } from './utils.js';
 
-const EXPORT_FORMATS = ['json', 'paper_json', 'print_payload'];
+const INLINE_EXPORT_FORMATS = ['json', 'paper_json', 'print_payload'];
+const BINARY_EXPORT_FORMATS = ['pdf', 'docx'];
+const EXPORT_FORMATS = [...INLINE_EXPORT_FORMATS, ...BINARY_EXPORT_FORMATS];
 const EXPORT_TYPES = ['questions', 'paper', 'daily_practice'];
 const SCOPE_TYPES = ['collection', 'entry', 'content_node'];
 const SUB_QUESTION_ANSWER_KEYS = new Set([
@@ -77,6 +79,44 @@ interface ExportQuestionRow {
   sourceHash: string | null;
 }
 
+export interface BuiltExportPayload {
+  _tikuExport: string;
+  jobId: string;
+  tenantId: string;
+  exportedAt: string;
+  exportType: string;
+  format: string;
+  title: string;
+  scope: ExportScope;
+  options: Record<string, unknown>;
+  summary: {
+    questionCount: number;
+    sectionCount: number;
+    totalScore: number;
+    durationMinutes: number | null;
+  };
+  sections: Array<{
+    key: string;
+    title: string;
+    questions: ReturnType<typeof formatQuestion>[];
+    totalScore: number;
+    questionCount: number;
+  }>;
+  questions: ReturnType<typeof formatQuestion>[];
+  files: Array<{
+    filename: string;
+    mimeType: string;
+    encoding: 'base64';
+    contentBase64: string;
+  }>;
+  renderHints: {
+    pdfLayout: string;
+    pageSize: string;
+    answerPlacement: string;
+    frontendRenderer: string;
+  };
+}
+
 function choose(value: unknown, allowed: string[], fallback: string, code: string) {
   const candidate = nullableString(value) || fallback;
   if (!allowed.includes(candidate)) {
@@ -99,6 +139,10 @@ function exportLimit(value: unknown) {
   return Math.max(1, Math.min(Math.trunc(parsed), 5000));
 }
 
+function isBinaryExportFormat(format: string) {
+  return BINARY_EXPORT_FORMATS.includes(format);
+}
+
 function jsonArray(value: unknown) {
   return Array.isArray(value) ? value : [];
 }
@@ -199,7 +243,7 @@ function buildExportPayload(input: {
   questions: ReturnType<typeof formatQuestion>[];
   options: Record<string, unknown>;
   createdAt: string;
-}) {
+}): BuiltExportPayload {
   const sections = groupBySection(input.questions);
   const title = typeof input.options.title === 'string' && input.options.title.trim()
     ? input.options.title.trim()
@@ -424,6 +468,138 @@ async function loadQuestions(client: pg.PoolClient, tenantId: string, scopeType:
   return result.rows;
 }
 
+async function createExportJob(input: {
+  client: pg.PoolClient;
+  auth: TenantContentAuth;
+  scope: ExportScope;
+  scopeType: ExportScope['scopeType'];
+  scopeId: string;
+  exportType: string;
+  format: string;
+  includeAnswers: boolean;
+  includeExplanations: boolean;
+  includeVideoRefs: boolean;
+  questionCount: number;
+  options: Record<string, unknown>;
+  outputMetadata: Record<string, unknown>;
+  status: 'pending' | 'completed';
+  finishedAt?: boolean;
+}) {
+  const job = await input.client.query<{
+    id: string;
+    createdAt: string;
+  }>(
+    `
+    insert into public.content_export_jobs (
+      tenant_id, created_by, export_type, format, scope_type, scope_id,
+      status, include_answers, include_explanations, include_video_refs,
+      question_count, summary, options, output_metadata, finished_at
+    )
+    values (
+      $1, $2, $3, $4, $5, $6,
+      $7, $8, $9, $10,
+      $11, $12::jsonb, $13::jsonb, $14::jsonb,
+      case when $15::boolean then now() else null end
+    )
+    returning id, created_at as "createdAt"
+    `,
+    [
+      input.auth.tenantId,
+      input.auth.userId,
+      input.exportType,
+      input.format,
+      input.scopeType,
+      input.scopeId,
+      input.status,
+      input.includeAnswers,
+      input.includeExplanations,
+      input.includeVideoRefs,
+      input.questionCount,
+      JSON.stringify({
+        questionCount: input.questionCount,
+        scopeName: input.scope.collectionName || input.scope.contentNodeName || input.scope.entryName,
+        format: input.format,
+        exportType: input.exportType,
+      }),
+      jsonObjectValue(input.options),
+      JSON.stringify(input.outputMetadata),
+      input.finishedAt === true,
+    ],
+  );
+  return job.rows[0];
+}
+
+export async function buildExportPayloadForJob(input: {
+  client: pg.PoolClient;
+  tenantId: string;
+  jobId: string;
+  createdAt?: string;
+}): Promise<BuiltExportPayload> {
+  const job = await input.client.query<{
+    id: string;
+    tenantId: string;
+    createdBy: string | null;
+    exportType: string;
+    format: string;
+    scopeType: ExportScope['scopeType'];
+    scopeId: string;
+    includeAnswers: boolean;
+    includeExplanations: boolean;
+    includeVideoRefs: boolean;
+    options: Record<string, unknown>;
+    outputMetadata: Record<string, unknown>;
+    createdAt: string;
+  }>(
+    `
+    select id, tenant_id as "tenantId", created_by as "createdBy",
+           export_type as "exportType", format,
+           scope_type as "scopeType", scope_id as "scopeId",
+           include_answers as "includeAnswers",
+           include_explanations as "includeExplanations",
+           include_video_refs as "includeVideoRefs",
+           options, output_metadata as "outputMetadata",
+           created_at as "createdAt"
+    from public.content_export_jobs
+    where tenant_id = $1 and id = $2
+    limit 1
+    `,
+    [input.tenantId, input.jobId],
+  );
+  const row = job.rows[0];
+  if (!row) throw new HttpError(404, 'Export job not found', 'EXPORT_JOB_NOT_FOUND');
+
+  const scope = await loadScope(input.client, row.tenantId, row.scopeType, row.scopeId);
+  const maxQuestionLimit = Number(row.outputMetadata?.maxQuestionLimit ?? row.options?.limit ?? 1000);
+  const limit = exportLimit(maxQuestionLimit);
+  const rows = await loadQuestions(input.client, row.tenantId, row.scopeType, row.scopeId, limit);
+  const questions = rows.map(question => formatQuestion(
+    question,
+    row.includeAnswers,
+    row.includeExplanations,
+    row.includeVideoRefs,
+  ));
+
+  return buildExportPayload({
+    auth: {
+      tenantId: row.tenantId,
+      userId: row.createdBy || row.tenantId,
+      role: 'system_worker',
+      permissions: { 'content:*': true },
+      templatePermissions: {},
+    },
+    jobId: row.id,
+    scope,
+    exportType: row.exportType,
+    format: row.format,
+    includeAnswers: row.includeAnswers,
+    includeExplanations: row.includeExplanations,
+    includeVideoRefs: row.includeVideoRefs,
+    questions,
+    options: row.options || {},
+    createdAt: input.createdAt || row.createdAt || new Date().toISOString(),
+  });
+}
+
 export async function createQuestionExportRoute(ctx: RequestContext) {
   const auth = await requireTenantContentEditor(ctx);
   const body = await readJsonBody(ctx);
@@ -457,64 +633,47 @@ export async function createQuestionExportRoute(ctx: RequestContext) {
       createdAt,
     };
     const outputHash = contentHash(prePayload);
+    const binaryExport = isBinaryExportFormat(format);
 
-    const job = await client.query<{
-      id: string;
-      createdAt: string;
-    }>(
-      `
-      insert into public.content_export_jobs (
-        tenant_id, created_by, export_type, format, scope_type, scope_id,
-        status, include_answers, include_explanations, include_video_refs,
-        question_count, summary, options, output_metadata, finished_at
-      )
-      values (
-        $1, $2, $3, $4, $5, $6,
-        'completed', $7, $8, $9,
-        $10, $11::jsonb, $12::jsonb, $13::jsonb, now()
-      )
-      returning id, created_at as "createdAt"
-      `,
-      [
-        auth.tenantId,
-        auth.userId,
-        exportType,
-        format,
-        scopeType,
-        scopeId,
-        includeAnswers,
-        includeExplanations,
-        includeVideoRefs,
-        questions.length,
-        JSON.stringify({
-          questionCount: questions.length,
-          scopeName: scope.collectionName || scope.contentNodeName || scope.entryName,
-          format,
-          exportType,
-        }),
-        jsonObjectValue(options),
-        JSON.stringify({
-          outputHash,
-          delivery: 'inline_payload',
-          maxQuestionLimit: limit,
-        }),
-      ],
-    );
-
-    const payload = buildExportPayload({
+    const job = await createExportJob({
+      client,
       auth,
-      jobId: job.rows[0].id,
       scope,
+      scopeType,
+      scopeId,
       exportType,
       format,
       includeAnswers,
       includeExplanations,
       includeVideoRefs,
-      questions,
-      options,
-      createdAt,
+      questionCount: questions.length,
+      options: { ...options, limit },
+      outputMetadata: {
+        outputHash,
+        delivery: binaryExport ? 'content_asset_pending' : 'inline_payload',
+        maxQuestionLimit: limit,
+        renderer: binaryExport ? 'worker:exports' : 'api:inline-json',
+      },
+      status: binaryExport ? 'pending' : 'completed',
+      finishedAt: !binaryExport,
     });
 
+    const payload = binaryExport
+      ? null
+      : buildExportPayload({
+        auth,
+        jobId: job.id,
+        scope,
+        exportType,
+        format,
+        includeAnswers,
+        includeExplanations,
+        includeVideoRefs,
+        questions,
+        options,
+        createdAt: job.createdAt || createdAt,
+      });
+
     await client.query(
       `
       insert into public.audit_logs (tenant_id, actor_user_id, action, target_type, target_id, details)
@@ -523,7 +682,7 @@ export async function createQuestionExportRoute(ctx: RequestContext) {
       [
         auth.tenantId,
         auth.userId,
-        job.rows[0].id,
+        job.id,
         JSON.stringify({
           scopeType,
           scopeId,
@@ -534,15 +693,16 @@ export async function createQuestionExportRoute(ctx: RequestContext) {
           includeExplanations,
           includeVideoRefs,
           outputHash,
+          delivery: binaryExport ? 'content_asset_pending' : 'inline_payload',
         }),
       ],
     );
 
     return {
       job: {
-        id: job.rows[0].id,
-        status: 'completed',
-        createdAt: job.rows[0].createdAt,
+        id: job.id,
+        status: binaryExport ? 'pending' : 'completed',
+        createdAt: job.createdAt,
         questionCount: questions.length,
         outputHash,
       },
@@ -580,7 +740,9 @@ export async function questionExportJobsRoute(ctx: RequestContext) {
            include_explanations as "includeExplanations",
            include_video_refs as "includeVideoRefs", question_count as "questionCount",
            summary, options, output_metadata as "outputMetadata",
-           error_message as "errorMessage", finished_at as "finishedAt",
+           asset_id as "assetId", error_message as "errorMessage",
+           started_at as "startedAt", finished_at as "finishedAt",
+           attempt_count as "attemptCount", max_attempts as "maxAttempts",
            created_by as "createdBy", created_at as "createdAt", updated_at as "updatedAt"
     from public.content_export_jobs
     where ${filters.join(' and ')}
diff --git a/apps/worker/package.json b/apps/worker/package.json
index 0e6263af..0c677910 100644
--- a/apps/worker/package.json
+++ b/apps/worker/package.json
@@ -12,15 +12,19 @@
     "commerce:once": "tsx src/index.ts --once --job commerce",
     "assets:once": "tsx src/index.ts --once --job assets",
     "imports:once": "tsx src/index.ts --once --job imports",
-    "public-banks:once": "tsx src/index.ts --once --job public-banks"
+    "public-banks:once": "tsx src/index.ts --once --job public-banks",
+    "exports:once": "tsx src/index.ts --once --job exports"
   },
   "dependencies": {
     "@supabase/storage-js": "^2.108.2",
     "ali-oss": "^6.23.0",
+    "docx": "^9.7.1",
+    "pdfkit": "^0.19.1",
     "pg": "^8.16.3"
   },
   "devDependencies": {
     "@types/node": "^24.0.4",
+    "@types/pdfkit": "^0.17.6",
     "@types/pg": "^8.15.4",
     "tsx": "^4.20.3",
     "typescript": "^5.8.3"
diff --git a/apps/worker/src/config.ts b/apps/worker/src/config.ts
index 6f801a85..b0f48b58 100644
--- a/apps/worker/src/config.ts
+++ b/apps/worker/src/config.ts
@@ -24,6 +24,14 @@ export interface WorkerConfig {
   publicBankSyncCopyLimit: number;
   publicBankSyncWorkerId: string;
   publicBankSyncClaimStaleSeconds: number;
+  exportBatchSize: number;
+  exportWorkerId: string;
+  exportBackoffSeconds: number[];
+  exportLocalStorageRoot: string;
+  exportPdfFontPath: string;
+  storageDefaultProvider: string;
+  storageDefaultBucket: string;
+  storagePublicBaseUrl: string;
   storageMaxUploadBytes: number;
   storageAllowedMimePrefixes: string[];
   storageAllowedMimeTypes: string[];
@@ -69,6 +77,16 @@ export const config: WorkerConfig = {
   publicBankSyncCopyLimit: envNumber('WORKER_PUBLIC_BANK_SYNC_COPY_LIMIT', 1000),
   publicBankSyncWorkerId: envString('WORKER_PUBLIC_BANK_SYNC_ID', `public-banks-${process.pid}`),
   publicBankSyncClaimStaleSeconds: envNumber('WORKER_PUBLIC_BANK_SYNC_CLAIM_STALE_SECONDS', 15 * 60),
+  exportBatchSize: envNumber('WORKER_EXPORT_BATCH_SIZE', 5),
+  exportWorkerId: envString('WORKER_EXPORT_ID', `exports-${process.pid}`),
+  exportBackoffSeconds: envList('WORKER_EXPORT_BACKOFF_SECONDS', '30,120,600,1800')
+    .map((value: string) => Number(value))
+    .filter((value: number) => Number.isFinite(value) && value > 0),
+  exportLocalStorageRoot: envString('EXPORT_LOCAL_STORAGE_ROOT', '.local-storage'),
+  exportPdfFontPath: envString('EXPORT_PDF_FONT_PATH', ''),
+  storageDefaultProvider: envString('STORAGE_DEFAULT_PROVIDER', 'local_dev'),
+  storageDefaultBucket: envString('STORAGE_DEFAULT_BUCKET', 'tenant-assets'),
+  storagePublicBaseUrl: envString('STORAGE_PUBLIC_BASE_URL', ''),
   storageMaxUploadBytes: envNumber('STORAGE_MAX_UPLOAD_BYTES', 1024 * 1024 * 500),
   storageAllowedMimePrefixes: envList('STORAGE_ALLOWED_MIME_PREFIXES', 'image/,video/,audio/'),
   storageAllowedMimeTypes: envList(
diff --git a/apps/worker/src/index.ts b/apps/worker/src/index.ts
index 358ac566..3510c91b 100644
--- a/apps/worker/src/index.ts
+++ b/apps/worker/src/index.ts
@@ -61,6 +61,17 @@ async function runOnce() {
     );
     return;
   }
+  if (job === 'exports') {
+    const { closeExportExecutorPool, processExportBatch } = await import('./jobs/exports.js');
+    extraClosers.add(closeExportExecutorPool);
+    const result = await processExportBatch();
+    console.log(
+      `[worker] exports batch processed=${result.processed}`
+        + ` completed=${result.completed} failed=${result.failed}`
+        + ` retrying=${result.retrying} skipped=${result.skipped}`,
+    );
+    return;
+  }
   throw new Error(`Unsupported worker job: ${job}`);
 }
 
diff --git a/apps/worker/src/jobs/export-renderer.ts b/apps/worker/src/jobs/export-renderer.ts
new file mode 100644
index 00000000..3efb4820
--- /dev/null
+++ b/apps/worker/src/jobs/export-renderer.ts
@@ -0,0 +1,340 @@
+import fs from 'node:fs';
+import PDFDocument from 'pdfkit';
+import {
+  AlignmentType,
+  Document,
+  Footer,
+  Header,
+  HeadingLevel,
+  Packer,
+  PageNumber,
+  Paragraph,
+  TextRun,
+} from 'docx';
+import type { BuiltExportPayload } from '../../../api/src/features/tenant-content/exports.js';
+import { config } from '../config.js';
+
+interface RenderResult {
+  body: Buffer;
+  mimeType: string;
+  extension: 'pdf' | 'docx';
+}
+
+type ExportQuestion = BuiltExportPayload['questions'][number];
+
+const PDF_FONT_CANDIDATES = [
+  config.exportPdfFontPath,
+  'C:\\Windows\\Fonts\\NotoSansSC-VF.ttf',
+  'C:\\Windows\\Fonts\\msyh.ttc',
+  'C:\\Windows\\Fonts\\simhei.ttf',
+  '/usr/share/fonts/opentype/noto/NotoSansCJK-Regular.ttc',
+  '/usr/share/fonts/truetype/noto/NotoSansCJK-Regular.ttc',
+  '/usr/share/fonts/truetype/wqy/wqy-microhei.ttc',
+].filter(Boolean);
+
+function plainText(value: unknown) {
+  return String(value ?? '')
+    .replace(/!\[([^\]]*)\]\(([^)]+)\)/g, ' [图片: $1 $2] ')
+    .replace(/\[([^\]]+)\]\(([^)]+)\)/g, '$1')
+    .replace(/<br\s*\/?>/gi, '\n')
+    .replace(/<\/(p|div|li|tr|h[1-6])>/gi, '\n')
+    .replace(/<[^>]+>/g, '')
+    .replace(/&nbsp;/g, ' ')
+    .replace(/&lt;/g, '<')
+    .replace(/&gt;/g, '>')
+    .replace(/&amp;/g, '&')
+    .replace(/\r\n/g, '\n')
+    .replace(/\n{3,}/g, '\n\n')
+    .trim();
+}
+
+function optionText(option: unknown, index: number) {
+  const prefix = `${String.fromCharCode(65 + index)}. `;
+  if (option && typeof option === 'object' && !Array.isArray(option)) {
+    const object = option as Record<string, unknown>;
+    return prefix + plainText(object.text ?? object.label ?? object.content ?? JSON.stringify(object));
+  }
+  return prefix + plainText(option);
+}
+
+function answerText(question: ExportQuestion) {
+  const answers: string[] = [];
+  if (typeof question.correctOptionIndex === 'number') {
+    answers.push(String.fromCharCode(65 + question.correctOptionIndex));
+  }
+  if (Array.isArray(question.correctOptionIndices) && question.correctOptionIndices.length) {
+    answers.push(question.correctOptionIndices.map(item => String.fromCharCode(65 + Number(item))).join(', '));
+  }
+  if (typeof question.answerText === 'string' && question.answerText.trim()) {
+    answers.push(plainText(question.answerText));
+  }
+  return Array.from(new Set(answers.filter(Boolean))).join('；');
+}
+
+function subQuestionLines(question: ExportQuestion, includeAnswers: boolean, includeExplanations: boolean) {
+  const subQuestions = Array.isArray(question.subQuestions) ? question.subQuestions : [];
+  const lines: string[] = [];
+  subQuestions.forEach((item, index) => {
+    if (!item || typeof item !== 'object' || Array.isArray(item)) return;
+    const object = item as Record<string, unknown>;
+    lines.push(`(${index + 1}) ${plainText(object.content ?? object.stem ?? object.title ?? '')}`);
+    const options = Array.isArray(object.options) ? object.options : [];
+    options.forEach((option, optionIndex) => {
+      lines.push(`    ${optionText(option, optionIndex)}`);
+    });
+    if (includeAnswers) {
+      const answer = plainText(object.answerText ?? object.answer_text ?? object.answer ?? object.correctAnswer ?? object.referenceAnswer ?? '');
+      const optionAnswer = Array.isArray(object.correctOptionIndices)
+        ? object.correctOptionIndices.map(value => String.fromCharCode(65 + Number(value))).join(', ')
+        : typeof object.correctOptionIndex === 'number'
+          ? String.fromCharCode(65 + object.correctOptionIndex)
+          : '';
+      if (answer || optionAnswer) lines.push(`    答案：${answer || optionAnswer}`);
+    }
+    if (includeExplanations) {
+      const explanation = plainText(object.explanation ?? object.analysis ?? '');
+      if (explanation) lines.push(`    解析：${explanation}`);
+    }
+  });
+  return lines;
+}
+
+function boolOption(payload: BuiltExportPayload, key: string) {
+  return payload.options && typeof payload.options[key] === 'boolean' ? payload.options[key] === true : false;
+}
+
+function watermarkText(payload: BuiltExportPayload) {
+  const raw = payload.options?.watermarkText;
+  return typeof raw === 'string' && raw.trim() ? raw.trim().slice(0, 80) : '';
+}
+
+function findPdfFont() {
+  return PDF_FONT_CANDIDATES.find(candidate => fs.existsSync(candidate));
+}
+
+function collectPdf(doc: PDFKit.PDFDocument) {
+  return new Promise<Buffer>((resolve, reject) => {
+    const chunks: Buffer[] = [];
+    doc.on('data', chunk => chunks.push(Buffer.from(chunk)));
+    doc.on('end', () => resolve(Buffer.concat(chunks)));
+    doc.on('error', reject);
+  });
+}
+
+function renderPdfWatermark(doc: PDFKit.PDFDocument, text: string) {
+  if (!text) return;
+  const width = doc.page.width;
+  const height = doc.page.height;
+  doc.save();
+  doc.rotate(-35, { origin: [width / 2, height / 2] });
+  doc.opacity(0.08);
+  doc.fillColor('#64748b');
+  doc.fontSize(36);
+  doc.text(text, 40, height / 2 - 40, { align: 'center', width: width - 80 });
+  doc.opacity(1);
+  doc.restore();
+  doc.fillColor('#111827');
+}
+
+function writePdfQuestion(doc: PDFKit.PDFDocument, question: ExportQuestion, index: number, payload: BuiltExportPayload) {
+  const includeAnswers = boolOption(payload, 'includeAnswers');
+  const includeExplanations = boolOption(payload, 'includeExplanations');
+  const typeLabel = typeof question.typeLabel === 'string' && question.typeLabel ? ` [${question.typeLabel}]` : '';
+  const score = typeof question.score === 'number' ? ` (${question.score}分)` : '';
+  doc.moveDown(0.4);
+  doc.fontSize(10).fillColor('#111827').text(`${index + 1}. ${plainText(question.content)}${typeLabel}${score}`, {
+    lineGap: 3,
+  });
+  const options = Array.isArray(question.options) ? question.options : [];
+  options.forEach((option, optionIndex) => {
+    doc.fontSize(9).fillColor('#374151').text(optionText(option, optionIndex), { indent: 16, lineGap: 2 });
+  });
+  for (const line of subQuestionLines(question, includeAnswers, includeExplanations)) {
+    doc.fontSize(9).fillColor('#374151').text(line, { indent: line.startsWith('(') ? 16 : 26, lineGap: 2 });
+  }
+  const answer = includeAnswers ? answerText(question) : '';
+  if (answer) {
+    doc.fontSize(9).fillColor('#0f766e').text(`答案：${answer}`, { indent: 16, lineGap: 2 });
+  }
+  if (includeExplanations && typeof question.explanation === 'string' && question.explanation.trim()) {
+    doc.fontSize(9).fillColor('#475569').text(`解析：${plainText(question.explanation)}`, { indent: 16, lineGap: 2 });
+  }
+}
+
+export async function renderExportPdf(payload: BuiltExportPayload): Promise<RenderResult> {
+  const doc = new PDFDocument({ size: 'A4', margin: 48, bufferPages: true, autoFirstPage: true });
+  const done = collectPdf(doc);
+  const fontPath = findPdfFont();
+  if (fontPath) {
+    doc.registerFont('TikuCjk', fontPath);
+    doc.font('TikuCjk');
+  }
+
+  const watermark = watermarkText(payload);
+  renderPdfWatermark(doc, watermark);
+  doc.on('pageAdded', () => {
+    if (fontPath) doc.font('TikuCjk');
+    renderPdfWatermark(doc, watermark);
+  });
+
+  doc.fontSize(18).fillColor('#0f172a').text(payload.title || '题库导出', { align: 'center' });
+  doc.moveDown(0.4);
+  doc.fontSize(9).fillColor('#64748b').text(
+    `导出时间：${payload.exportedAt}    题量：${payload.summary.questionCount}    总分：${payload.summary.totalScore}`,
+    { align: 'center' },
+  );
+  if (watermark) {
+    doc.moveDown(0.3);
+    doc.fontSize(8).fillColor('#94a3b8').text(`水印：${watermark}`, { align: 'center' });
+  }
+
+  let questionIndex = 0;
+  for (const section of payload.sections) {
+    doc.moveDown(1);
+    doc.fontSize(13).fillColor('#1e3a8a').text(section.title || section.key, { underline: true });
+    for (const question of section.questions) {
+      writePdfQuestion(doc, question, questionIndex, payload);
+      questionIndex += 1;
+    }
+  }
+
+  const pages = doc.bufferedPageRange();
+  for (let i = 0; i < pages.count; i += 1) {
+    doc.switchToPage(i);
+    if (fontPath) doc.font('TikuCjk');
+    doc.fontSize(8).fillColor('#94a3b8').text(
+      `${payload.title || '题库导出'} · 第 ${i + 1} / ${pages.count} 页`,
+      48,
+      doc.page.height - 34,
+      { align: 'center', width: doc.page.width - 96 },
+    );
+  }
+  doc.end();
+  return {
+    body: await done,
+    mimeType: 'application/pdf',
+    extension: 'pdf',
+  };
+}
+
+function docxParagraph(text: string, options: { heading?: (typeof HeadingLevel)[keyof typeof HeadingLevel]; bold?: boolean } = {}) {
+  return new Paragraph({
+    heading: options.heading,
+    spacing: { after: 140 },
+    children: [
+      new TextRun({
+        text,
+        bold: options.bold,
+        font: 'Microsoft YaHei',
+        size: options.heading ? 28 : 21,
+      }),
+    ],
+  });
+}
+
+function docxQuestionParagraphs(question: ExportQuestion, index: number, payload: BuiltExportPayload) {
+  const includeAnswers = boolOption(payload, 'includeAnswers');
+  const includeExplanations = boolOption(payload, 'includeExplanations');
+  const paragraphs: Paragraph[] = [];
+  const typeLabel = typeof question.typeLabel === 'string' && question.typeLabel ? ` [${question.typeLabel}]` : '';
+  const score = typeof question.score === 'number' ? ` (${question.score}分)` : '';
+  paragraphs.push(docxParagraph(`${index + 1}. ${plainText(question.content)}${typeLabel}${score}`, { bold: true }));
+  const options = Array.isArray(question.options) ? question.options : [];
+  options.forEach((option, optionIndex) => {
+    paragraphs.push(docxParagraph(optionText(option, optionIndex)));
+  });
+  for (const line of subQuestionLines(question, includeAnswers, includeExplanations)) {
+    paragraphs.push(docxParagraph(line));
+  }
+  const answer = includeAnswers ? answerText(question) : '';
+  if (answer) paragraphs.push(docxParagraph(`答案：${answer}`));
+  if (includeExplanations && typeof question.explanation === 'string' && question.explanation.trim()) {
+    paragraphs.push(docxParagraph(`解析：${plainText(question.explanation)}`));
+  }
+  return paragraphs;
+}
+
+export async function renderExportDocx(payload: BuiltExportPayload): Promise<RenderResult> {
+  const watermark = watermarkText(payload);
+  const children: Paragraph[] = [
+    new Paragraph({
+      heading: HeadingLevel.TITLE,
+      alignment: AlignmentType.CENTER,
+      spacing: { after: 180 },
+      children: [new TextRun({ text: payload.title || '题库导出', bold: true, font: 'Microsoft YaHei', size: 36 })],
+    }),
+    new Paragraph({
+      alignment: AlignmentType.CENTER,
+      spacing: { after: 180 },
+      children: [
+        new TextRun({
+          text: `导出时间：${payload.exportedAt}    题量：${payload.summary.questionCount}    总分：${payload.summary.totalScore}`,
+          color: '64748B',
+          font: 'Microsoft YaHei',
+          size: 19,
+        }),
+      ],
+    }),
+  ];
+  if (watermark) {
+    children.push(new Paragraph({
+      alignment: AlignmentType.CENTER,
+      spacing: { after: 180 },
+      children: [new TextRun({ text: `水印：${watermark}`, color: '94A3B8', font: 'Microsoft YaHei', size: 18 })],
+    }));
+  }
+  let questionIndex = 0;
+  for (const section of payload.sections) {
+    children.push(docxParagraph(section.title || section.key, { heading: HeadingLevel.HEADING_2, bold: true }));
+    for (const question of section.questions) {
+      children.push(...docxQuestionParagraphs(question, questionIndex, payload));
+      questionIndex += 1;
+    }
+  }
+
+  const doc = new Document({
+    creator: 'Tiku SaaS Export Worker',
+    title: payload.title || '题库导出',
+    description: 'Generated by Tiku SaaS export worker',
+    sections: [
+      {
+        headers: {
+          default: new Header({
+            children: [
+              new Paragraph({
+                alignment: AlignmentType.CENTER,
+                children: [
+                  new TextRun({
+                    text: watermark || payload.title || '题库导出',
+                    color: 'CBD5E1',
+                    font: 'Microsoft YaHei',
+                    size: 18,
+                  }),
+                ],
+              }),
+            ],
+          }),
+        },
+        footers: {
+          default: new Footer({
+            children: [
+              new Paragraph({
+                alignment: AlignmentType.CENTER,
+                children: [
+                  new TextRun({ children: ['第 ', PageNumber.CURRENT, ' / ', PageNumber.TOTAL_PAGES, ' 页'], font: 'Microsoft YaHei', size: 18 }),
+                ],
+              }),
+            ],
+          }),
+        },
+        children,
+      },
+    ],
+  });
+
+  return {
+    body: await Packer.toBuffer(doc),
+    mimeType: 'application/vnd.openxmlformats-officedocument.wordprocessingml.document',
+    extension: 'docx',
+  };
+}
diff --git a/apps/worker/src/jobs/exports.ts b/apps/worker/src/jobs/exports.ts
new file mode 100644
index 00000000..a3cb56ae
--- /dev/null
+++ b/apps/worker/src/jobs/exports.ts
@@ -0,0 +1,496 @@
+import path from 'node:path';
+import { randomUUID } from 'node:crypto';
+import type pg from 'pg';
+import { pool } from '../db.js';
+import { config } from '../config.js';
+import { closePool as closeApiExportPool } from '../../../api/src/core/db.js';
+import { buildExportPayloadForJob } from '../../../api/src/features/tenant-content/exports.js';
+import { normalizeStorageProvider, putStorageObject, storageBucket, validateObjectKey } from '../storage.js';
+import { renderExportDocx, renderExportPdf } from './export-renderer.js';
+
+interface ExportJobRow {
+  id: string;
+  tenantId: string;
+  createdBy: string | null;
+  exportType: string;
+  format: 'pdf' | 'docx';
+  scopeType: string;
+  scopeId: string;
+  status: string;
+  attemptCount: number;
+  maxAttempts: number;
+  options: Record<string, unknown>;
+  outputMetadata: Record<string, unknown>;
+  createdAt: string;
+}
+
+interface ExportWorkerResult {
+  processed: number;
+  completed: number;
+  failed: number;
+  retrying: number;
+  skipped: number;
+}
+
+interface RenderedExport {
+  body: Buffer;
+  mimeType: string;
+  extension: 'pdf' | 'docx';
+}
+
+const ASSET_VISIBILITIES = new Set(['public', 'tenant', 'members', 'svip', 'private']);
+
+function objectValue(value: unknown): Record<string, unknown> {
+  return value && typeof value === 'object' && !Array.isArray(value) ? value as Record<string, unknown> : {};
+}
+
+function boolValue(value: unknown, fallback: boolean) {
+  return typeof value === 'boolean' ? value : fallback;
+}
+
+function stringValue(value: unknown, fallback = '') {
+  return typeof value === 'string' && value.trim() ? value.trim() : fallback;
+}
+
+function truncate(value: unknown, max = 1900) {
+  return String(value ?? '').slice(0, max);
+}
+
+function errorMessage(error: unknown) {
+  return error instanceof Error ? error.message : String(error);
+}
+
+function errorCode(error: unknown) {
+  return typeof error === 'object' && error !== null && 'code' in error
+    ? String((error as { code?: unknown }).code || 'EXPORT_WORKER_ERROR')
+    : 'EXPORT_WORKER_ERROR';
+}
+
+function backoffSeconds(attemptCount: number) {
+  const backoffs = config.exportBackoffSeconds.length ? config.exportBackoffSeconds : [30, 120, 600, 1800];
+  return backoffs[Math.min(Math.max(0, attemptCount - 1), backoffs.length - 1)];
+}
+
+function safeFileBase(value: string) {
+  const ascii = value
+    .normalize('NFKD')
+    .replace(/[^\w.-]+/g, '-')
+    .replace(/-+/g, '-')
+    .replace(/^-+|-+$/g, '')
+    .slice(0, 80);
+  return ascii || 'question-export';
+}
+
+function safeDisplayFileName(title: string, extension: string) {
+  return `${title.trim().replace(/[\\/:*?"<>|]+/g, '-').replace(/\s+/g, '-').slice(0, 120) || '题库导出'}.${extension}`;
+}
+
+function objectKeyFor(job: ExportJobRow, title: string, extension: string) {
+  const created = new Date(job.createdAt || Date.now());
+  const yyyy = Number.isFinite(created.getTime()) ? created.getUTCFullYear() : new Date().getUTCFullYear();
+  const mm = String((Number.isFinite(created.getTime()) ? created.getUTCMonth() : new Date().getUTCMonth()) + 1).padStart(2, '0');
+  return validateObjectKey(
+    job.tenantId,
+    `${job.tenantId}/exports/${yyyy}/${mm}/${job.id}-${safeFileBase(title)}.${extension}`,
+  );
+}
+
+function assetVisibility(options: Record<string, unknown>) {
+  const requested = stringValue(options.assetVisibility, boolValue(options.publishToAssets, false) ? 'tenant' : 'private');
+  return ASSET_VISIBILITIES.has(requested) ? requested : 'private';
+}
+
+function assetStatus(options: Record<string, unknown>) {
+  return boolValue(options.publishToAssets, false) ? 'active' : 'active';
+}
+
+function accessRules(options: Record<string, unknown>) {
+  return objectValue(options.assetAccessRules);
+}
+
+async function claimExportJobs() {
+  const client = await pool.connect();
+  try {
+    await client.query('begin');
+    const result = await client.query<ExportJobRow>(
+      `
+      select id, tenant_id as "tenantId", created_by as "createdBy",
+             export_type as "exportType", format, scope_type as "scopeType",
+             scope_id as "scopeId", status, attempt_count as "attemptCount",
+             max_attempts as "maxAttempts", options,
+             output_metadata as "outputMetadata", created_at as "createdAt"
+      from public.content_export_jobs
+      where status = 'pending'
+        and format in ('pdf', 'docx')
+        and attempt_count < max_attempts
+        and (next_attempt_at is null or next_attempt_at <= now())
+      order by created_at asc
+      limit $1
+      for update skip locked
+      `,
+      [config.exportBatchSize],
+    );
+
+    const ids = result.rows.map(row => row.id);
+    if (ids.length) {
+      await client.query(
+        `
+        update public.content_export_jobs
+        set status = 'rendering',
+            locked_at = now(),
+            locked_by = $2,
+            attempt_count = attempt_count + 1,
+            started_at = coalesce(started_at, now()),
+            error_message = null,
+            updated_at = now()
+        where id = any($1::uuid[])
+        `,
+        [ids, config.exportWorkerId],
+      );
+    }
+    await client.query('commit');
+    return result.rows;
+  } catch (error) {
+    await client.query('rollback');
+    throw error;
+  } finally {
+    client.release();
+  }
+}
+
+async function renderJob(job: ExportJobRow): Promise<{ rendered: RenderedExport; title: string; questionCount: number }> {
+  const client = await pool.connect();
+  try {
+    const payload = await buildExportPayloadForJob({
+      client,
+      tenantId: job.tenantId,
+      jobId: job.id,
+      createdAt: new Date().toISOString(),
+    });
+    const rendered = job.format === 'pdf'
+      ? await renderExportPdf(payload)
+      : await renderExportDocx(payload);
+    return {
+      rendered,
+      title: payload.title || '题库导出',
+      questionCount: payload.summary.questionCount,
+    };
+  } finally {
+    client.release();
+  }
+}
+
+async function insertExportAsset(
+  client: pg.PoolClient,
+  job: ExportJobRow,
+  input: {
+    title: string;
+    fileName: string;
+    extension: 'pdf' | 'docx';
+    mimeType: string;
+    bucket: string;
+    objectKey: string;
+    sizeBytes: number;
+    checksumSha256: string;
+    uploadDetails: Record<string, unknown>;
+  },
+) {
+  const visibility = assetVisibility(job.options || {});
+  const status = assetStatus(job.options || {});
+  const assetType = input.extension === 'pdf' ? 'pdf' : 'document';
+  const asset = await client.query<{ id: string }>(
+    `
+    insert into public.content_assets (
+      tenant_id, asset_key, title, category, description, file_name,
+      asset_type, storage_provider, bucket, object_key, mime_type,
+      file_size_bytes, checksum_sha256, upload_status, verified_at,
+      verified_size_bytes, verified_checksum_sha256, verification_details,
+      preview_status, visibility, is_public, status, access_rules,
+      metadata, created_by, updated_by, source
+    )
+    values (
+      $1, $2, $3, $4, $5, $6,
+      $7, $8, $9, $10, $11,
+      $12, $13, 'verified', now(),
+      $12, $13, $14::jsonb,
+      $15, $16, $17, $18, $19::jsonb,
+      $20::jsonb, $21, $21, 'content_export_worker'
+    )
+    returning id
+    `,
+    [
+      job.tenantId,
+      `content-export:${job.id}`,
+      input.title,
+      '题库导出',
+      `由题库导出任务 ${job.id} 自动生成`,
+      input.fileName,
+      assetType,
+      normalizeStorageProvider(String(job.outputMetadata?.storageProvider || config.storageDefaultProvider)),
+      input.bucket,
+      input.objectKey,
+      input.mimeType,
+      input.sizeBytes,
+      input.checksumSha256,
+      JSON.stringify({
+        exportWorker: {
+          ...input.uploadDetails,
+          workerId: config.exportWorkerId,
+          checkedAt: new Date().toISOString(),
+        },
+      }),
+      input.extension === 'pdf' ? 'ready' : 'none',
+      visibility,
+      visibility === 'public',
+      status,
+      JSON.stringify(accessRules(job.options || {})),
+      JSON.stringify({
+        exportJobId: job.id,
+        exportType: job.exportType,
+        format: job.format,
+        scopeType: job.scopeType,
+        scopeId: job.scopeId,
+        publishToAssets: boolValue(job.options?.publishToAssets, false),
+        watermarkText: stringValue(job.options?.watermarkText, ''),
+      }),
+      job.createdBy,
+    ],
+  );
+  return asset.rows[0].id;
+}
+
+async function markExportCompleted(job: ExportJobRow, input: {
+  assetId: string;
+  title: string;
+  fileName: string;
+  mimeType: string;
+  sizeBytes: number;
+  checksumSha256: string;
+  bucket: string;
+  objectKey: string;
+  questionCount: number;
+  uploadDetails: Record<string, unknown>;
+}) {
+  const client = await pool.connect();
+  try {
+    await client.query('begin');
+    await client.query(
+      `
+      update public.content_export_jobs
+      set status = 'completed',
+          asset_id = $3,
+          question_count = $4,
+          summary = coalesce(summary, '{}'::jsonb) || $5::jsonb,
+          output_metadata = coalesce(output_metadata, '{}'::jsonb) || $6::jsonb,
+          error_message = null,
+          locked_at = null,
+          locked_by = null,
+          next_attempt_at = null,
+          finished_at = now(),
+          updated_at = now()
+      where tenant_id = $1 and id = $2
+      `,
+      [
+        job.tenantId,
+        job.id,
+        input.assetId,
+        input.questionCount,
+        JSON.stringify({
+          questionCount: input.questionCount,
+          renderedTitle: input.title,
+        }),
+        JSON.stringify({
+          delivery: 'content_asset',
+          assetId: input.assetId,
+          fileName: input.fileName,
+          mimeType: input.mimeType,
+          sizeBytes: input.sizeBytes,
+          checksumSha256: input.checksumSha256,
+          bucket: input.bucket,
+          objectKey: input.objectKey,
+          renderer: 'worker:exports',
+          renderedAt: new Date().toISOString(),
+          upload: input.uploadDetails,
+        }),
+      ],
+    );
+    await client.query(
+      `
+      insert into public.audit_logs (tenant_id, actor_user_id, action, target_type, target_id, details)
+      values ($1, $2, 'content.questions.export_rendered', 'content_export_job', $3, $4::jsonb)
+      `,
+      [
+        job.tenantId,
+        job.createdBy,
+        job.id,
+        JSON.stringify({
+          assetId: input.assetId,
+          format: job.format,
+          mimeType: input.mimeType,
+          sizeBytes: input.sizeBytes,
+          checksumSha256: input.checksumSha256,
+          workerId: config.exportWorkerId,
+        }),
+      ],
+    );
+    await client.query('commit');
+  } catch (error) {
+    await client.query('rollback');
+    throw error;
+  } finally {
+    client.release();
+  }
+}
+
+async function markExportFailed(job: ExportJobRow, error: unknown) {
+  const nextAttempt = job.attemptCount + 1;
+  const willRetry = nextAttempt < job.maxAttempts;
+  const status = willRetry ? 'pending' : 'failed';
+  await pool.query(
+    `
+    update public.content_export_jobs
+    set status = $3,
+        error_message = $4,
+        output_metadata = coalesce(output_metadata, '{}'::jsonb) || $5::jsonb,
+        next_attempt_at = case when $6::boolean then now() + make_interval(secs => $7::integer) else null end,
+        locked_at = null,
+        locked_by = null,
+        finished_at = case when $3 = 'failed' then now() else finished_at end,
+        updated_at = now()
+    where tenant_id = $1 and id = $2
+    `,
+    [
+      job.tenantId,
+      job.id,
+      status,
+      truncate(errorMessage(error)),
+      JSON.stringify({
+        lastWorkerError: {
+          code: errorCode(error),
+          message: truncate(errorMessage(error)),
+          workerId: config.exportWorkerId,
+          failedAt: new Date().toISOString(),
+          nextAttempt,
+          maxAttempts: job.maxAttempts,
+          willRetry,
+        },
+      }),
+      willRetry,
+      backoffSeconds(nextAttempt),
+    ],
+  );
+
+  await pool.query(
+    `
+    insert into public.audit_logs (tenant_id, actor_user_id, action, target_type, target_id, details)
+    values ($1, $2, $3, 'content_export_job', $4, $5::jsonb)
+    `,
+    [
+      job.tenantId,
+      job.createdBy,
+      willRetry ? 'content.questions.export_retry_scheduled' : 'content.questions.export_failed',
+      job.id,
+      JSON.stringify({
+        code: errorCode(error),
+        message: truncate(errorMessage(error)),
+        workerId: config.exportWorkerId,
+        nextAttempt,
+        maxAttempts: job.maxAttempts,
+      }),
+    ],
+  );
+
+  return willRetry ? 'retrying' : 'failed';
+}
+
+async function processExportJob(job: ExportJobRow) {
+  const { rendered, title, questionCount } = await renderJob(job);
+  const extension = rendered.extension;
+  const bucket = storageBucket(String(job.outputMetadata?.storageBucket || ''));
+  const objectKey = objectKeyFor(job, title, extension);
+  const fileName = safeDisplayFileName(title, extension);
+  const upload = await putStorageObject({
+    tenantId: job.tenantId,
+    provider: normalizeStorageProvider(String(job.outputMetadata?.storageProvider || config.storageDefaultProvider)),
+    bucket,
+    objectKey,
+    body: rendered.body,
+    mimeType: rendered.mimeType,
+  });
+
+  const client = await pool.connect();
+  let assetId = '';
+  try {
+    await client.query('begin');
+    assetId = await insertExportAsset(client, job, {
+      title,
+      fileName,
+      extension,
+      mimeType: rendered.mimeType,
+      bucket,
+      objectKey,
+      sizeBytes: upload.sizeBytes,
+      checksumSha256: upload.checksumSha256,
+      uploadDetails: {
+        provider: upload.provider,
+        localPath: upload.localPath ? path.relative(process.cwd(), upload.localPath) : undefined,
+        etag: upload.etag,
+      },
+    });
+    await client.query('commit');
+  } catch (error) {
+    await client.query('rollback');
+    throw error;
+  } finally {
+    client.release();
+  }
+
+  await markExportCompleted(job, {
+    assetId,
+    title,
+    fileName,
+    mimeType: rendered.mimeType,
+    sizeBytes: upload.sizeBytes,
+    checksumSha256: upload.checksumSha256,
+    bucket,
+    objectKey,
+    questionCount,
+    uploadDetails: {
+      provider: upload.provider,
+      localPath: upload.localPath ? path.relative(process.cwd(), upload.localPath) : undefined,
+      etag: upload.etag,
+    },
+  });
+}
+
+export async function processExportBatch(): Promise<ExportWorkerResult> {
+  const jobs = await claimExportJobs();
+  const result: ExportWorkerResult = {
+    processed: jobs.length,
+    completed: 0,
+    failed: 0,
+    retrying: 0,
+    skipped: 0,
+  };
+
+  for (const job of jobs) {
+    try {
+      if (job.format !== 'pdf' && job.format !== 'docx') {
+        result.skipped += 1;
+        continue;
+      }
+      await processExportJob(job);
+      result.completed += 1;
+    } catch (error) {
+      const state = await markExportFailed(job, error);
+      if (state === 'retrying') result.retrying += 1;
+      else result.failed += 1;
+    }
+  }
+
+  return result;
+}
+
+export async function closeExportExecutorPool() {
+  await closeApiExportPool();
+}
diff --git a/apps/worker/src/storage.ts b/apps/worker/src/storage.ts
new file mode 100644
index 00000000..b457b7e4
--- /dev/null
+++ b/apps/worker/src/storage.ts
@@ -0,0 +1,246 @@
+import fs from 'node:fs/promises';
+import path from 'node:path';
+import crypto from 'node:crypto';
+import { StorageClient } from '@supabase/storage-js';
+import { config } from './config.js';
+
+export type StorageProviderName = 'external_url' | 'supabase_storage' | 'aliyun_oss' | 'tencent_cos' | 'qiniu_kodo' | 'local_dev';
+
+interface PutObjectInput {
+  tenantId: string;
+  provider: StorageProviderName;
+  bucket: string;
+  objectKey: string;
+  body: Buffer;
+  mimeType: string;
+}
+
+interface PutObjectResult {
+  provider: StorageProviderName;
+  bucket: string;
+  objectKey: string;
+  checksumSha256: string;
+  sizeBytes: number;
+  mimeType: string;
+  localPath?: string;
+  etag?: string | null;
+}
+
+const SAFE_OBJECT_KEY_RE = /^[A-Za-z0-9][A-Za-z0-9._~!$&'()+,;=@/-]{0,1023}$/;
+const SUPPORTED_UPLOAD_PROVIDERS = new Set<StorageProviderName>(['local_dev', 'supabase_storage', 'aliyun_oss', 'tencent_cos']);
+
+export function normalizeStorageProvider(value: string | null | undefined): StorageProviderName {
+  const provider = (value?.trim() || config.storageDefaultProvider || 'local_dev') as StorageProviderName;
+  if (!SUPPORTED_UPLOAD_PROVIDERS.has(provider)) {
+    throw new Error(`Unsupported export storage provider: ${provider}`);
+  }
+  return provider;
+}
+
+export function storageBucket(value: string | null | undefined) {
+  return value?.trim() || config.storageDefaultBucket || 'tenant-assets';
+}
+
+export function canonicalObjectKey(objectKey: string) {
+  return objectKey.replace(/^\/+/, '').replace(/\/{2,}/g, '/');
+}
+
+export function validateObjectKey(tenantId: string, objectKey: string) {
+  const clean = canonicalObjectKey(objectKey);
+  if (!clean || clean.includes('..') || clean.includes('\\') || clean.includes('%2f') || clean.includes('%2F')) {
+    throw new Error('Invalid objectKey');
+  }
+  if (!SAFE_OBJECT_KEY_RE.test(clean)) {
+    throw new Error('objectKey contains unsafe characters');
+  }
+  if (config.storageRequireTenantPrefix && !clean.startsWith(`${tenantId}/`)) {
+    throw new Error('objectKey must be scoped by tenantId prefix');
+  }
+  return clean;
+}
+
+function sha256(body: Buffer) {
+  return crypto.createHash('sha256').update(body).digest('hex');
+}
+
+function requireConfigured(condition: unknown, provider: StorageProviderName, missing: string) {
+  if (!condition) throw new Error(`${provider} is not configured: ${missing}`);
+}
+
+async function putLocalDev(input: PutObjectInput, checksumSha256: string): Promise<PutObjectResult> {
+  const root = path.resolve(process.cwd(), config.exportLocalStorageRoot || '.local-storage');
+  const bucketRoot = path.resolve(root, input.bucket);
+  const target = path.resolve(bucketRoot, input.objectKey);
+  if (!target.startsWith(bucketRoot + path.sep)) {
+    throw new Error('Resolved local object path escapes storage root');
+  }
+  await fs.mkdir(path.dirname(target), { recursive: true });
+  await fs.writeFile(target, input.body);
+  return {
+    provider: input.provider,
+    bucket: input.bucket,
+    objectKey: input.objectKey,
+    checksumSha256,
+    sizeBytes: input.body.byteLength,
+    mimeType: input.mimeType,
+    localPath: target,
+  };
+}
+
+async function putAliyunOss(input: PutObjectInput, checksumSha256: string): Promise<PutObjectResult> {
+  requireConfigured(config.aliyunOssAccessKeyId, 'aliyun_oss', 'ALIYUN_OSS_ACCESS_KEY_ID');
+  requireConfigured(config.aliyunOssAccessKeySecret, 'aliyun_oss', 'ALIYUN_OSS_ACCESS_KEY_SECRET');
+  requireConfigured(config.aliyunOssRegion || config.aliyunOssEndpoint, 'aliyun_oss', 'ALIYUN_OSS_REGION or ALIYUN_OSS_ENDPOINT');
+
+  const { default: OSS } = await import('ali-oss');
+  const client = new OSS({
+    region: config.aliyunOssRegion || undefined,
+    endpoint: config.aliyunOssEndpoint || undefined,
+    accessKeyId: config.aliyunOssAccessKeyId,
+    accessKeySecret: config.aliyunOssAccessKeySecret,
+    stsToken: config.aliyunOssStsToken || undefined,
+    bucket: input.bucket,
+    internal: config.aliyunOssInternal,
+    secure: true,
+  });
+  const response = await (client as unknown as {
+    put: (objectKey: string, body: Buffer, options: Record<string, unknown>) => Promise<{ res?: { headers?: Record<string, string> } }>;
+  }).put(input.objectKey, input.body, {
+    headers: {
+      'Content-Type': input.mimeType,
+      'x-oss-meta-sha256': checksumSha256,
+    },
+  });
+  return {
+    provider: input.provider,
+    bucket: input.bucket,
+    objectKey: input.objectKey,
+    checksumSha256,
+    sizeBytes: input.body.byteLength,
+    mimeType: input.mimeType,
+    etag: response.res?.headers?.etag || null,
+  };
+}
+
+async function putSupabaseStorage(input: PutObjectInput, checksumSha256: string): Promise<PutObjectResult> {
+  requireConfigured(config.supabaseStorageUrl, 'supabase_storage', 'SUPABASE_STORAGE_URL');
+  requireConfigured(config.supabaseStorageServiceKey, 'supabase_storage', 'SUPABASE_STORAGE_SERVICE_KEY');
+  const client = new StorageClient(config.supabaseStorageUrl.replace(/\/+$/, ''), {
+    apikey: config.supabaseStorageServiceKey,
+    authorization: `Bearer ${config.supabaseStorageServiceKey}`,
+  });
+  const response = await client.from(input.bucket).upload(input.objectKey, input.body, {
+    contentType: input.mimeType,
+    upsert: true,
+    duplex: 'half',
+    metadata: { sha256: checksumSha256 },
+  } as never);
+  if (response.error) {
+    throw new Error(response.error.message || 'Supabase Storage upload failed');
+  }
+  return {
+    provider: input.provider,
+    bucket: input.bucket,
+    objectKey: input.objectKey,
+    checksumSha256,
+    sizeBytes: input.body.byteLength,
+    mimeType: input.mimeType,
+  };
+}
+
+function hmacSha1Hex(key: string | Buffer, value: string) {
+  return crypto.createHmac('sha1', key).update(value).digest('hex');
+}
+
+function sha1Hex(value: string) {
+  return crypto.createHash('sha1').update(value).digest('hex');
+}
+
+function cosEncodePath(objectKey: string) {
+  return objectKey
+    .split('/')
+    .map(part => encodeURIComponent(part).replace(/[!'()*]/g, char => `%${char.charCodeAt(0).toString(16).toUpperCase()}`))
+    .join('/');
+}
+
+function cosHost(bucket: string) {
+  requireConfigured(config.tencentCosRegion, 'tencent_cos', 'TENCENT_COS_REGION');
+  const bucketWithAppId = config.tencentCosAppId && !bucket.endsWith(`-${config.tencentCosAppId}`)
+    ? `${bucket}-${config.tencentCosAppId}`
+    : bucket;
+  return `${bucketWithAppId}.cos.${config.tencentCosRegion}.myqcloud.com`;
+}
+
+function nowSeconds() {
+  return Math.floor(Date.now() / 1000);
+}
+
+async function putTencentCos(input: PutObjectInput, checksumSha256: string): Promise<PutObjectResult> {
+  requireConfigured(config.tencentCosSecretId, 'tencent_cos', 'TENCENT_COS_SECRET_ID');
+  requireConfigured(config.tencentCosSecretKey, 'tencent_cos', 'TENCENT_COS_SECRET_KEY');
+  const host = cosHost(input.bucket);
+  const start = nowSeconds();
+  const end = start + 300;
+  const keyTime = `${start};${end}`;
+  const pathname = `/${cosEncodePath(input.objectKey)}`;
+  const signedHeaders: Record<string, string> = {
+    host,
+    'content-type': input.mimeType,
+    'x-cos-meta-sha256': checksumSha256,
+  };
+  const headerKeys = Object.keys(signedHeaders).sort();
+  const headerList = headerKeys.join(';');
+  const httpHeaders = headerKeys
+    .map(key => `${encodeURIComponent(key)}=${encodeURIComponent(signedHeaders[key]).toLowerCase()}`)
+    .join('&');
+  const signedQuery: Record<string, string> = {};
+  if (config.tencentCosSecurityToken) signedQuery['x-cos-security-token'] = config.tencentCosSecurityToken;
+  const queryKeys = Object.keys(signedQuery).sort();
+  const urlParamList = queryKeys.join(';');
+  const httpParameters = queryKeys
+    .map(key => `${encodeURIComponent(key)}=${encodeURIComponent(signedQuery[key])}`)
+    .join('&');
+  const httpString = `put\n${pathname}\n${httpParameters}\n${httpHeaders}\n`;
+  const stringToSign = `sha1\n${keyTime}\n${sha1Hex(httpString)}\n`;
+  const signKey = hmacSha1Hex(config.tencentCosSecretKey, keyTime);
+  const signature = hmacSha1Hex(signKey, stringToSign);
+  const query = new URLSearchParams();
+  query.set('q-sign-algorithm', 'sha1');
+  query.set('q-ak', config.tencentCosSecretId);
+  query.set('q-sign-time', keyTime);
+  query.set('q-key-time', keyTime);
+  query.set('q-header-list', headerList);
+  query.set('q-url-param-list', urlParamList);
+  query.set('q-signature', signature);
+  for (const key of queryKeys) query.set(key, signedQuery[key]);
+
+  const response = await fetch(`https://${host}${pathname}?${query.toString()}`, {
+    method: 'PUT',
+    headers: signedHeaders,
+    body: input.body as unknown as BodyInit,
+  });
+  if (!response.ok) {
+    throw new Error(`Tencent COS upload failed: ${response.status}`);
+  }
+  return {
+    provider: input.provider,
+    bucket: input.bucket,
+    objectKey: input.objectKey,
+    checksumSha256,
+    sizeBytes: input.body.byteLength,
+    mimeType: input.mimeType,
+    etag: response.headers.get('etag'),
+  };
+}
+
+export async function putStorageObject(input: PutObjectInput): Promise<PutObjectResult> {
+  const provider = normalizeStorageProvider(input.provider);
+  const objectKey = validateObjectKey(input.tenantId, input.objectKey);
+  const normalized = { ...input, provider, objectKey };
+  const checksumSha256 = sha256(input.body);
+  if (provider === 'local_dev') return putLocalDev(normalized, checksumSha256);
+  if (provider === 'aliyun_oss') return putAliyunOss(normalized, checksumSha256);
+  if (provider === 'tencent_cos') return putTencentCos(normalized, checksumSha256);
+  if (provider === 'supabase_storage') return putSupabaseStorage(normalized, checksumSha256);
+  throw new Error(`${provider} does not support worker object upload`);
+}
diff --git a/docs/refactor/backend-capability-status.md b/docs/refactor/backend-capability-status.md
index 4227f7b7..5c78eee5 100644
--- a/docs/refactor/backend-capability-status.md
+++ b/docs/refactor/backend-capability-status.md
@@ -137,7 +137,7 @@
 | 数据看板聚合接口 | 可联调 | `GET /api/tenant-admin/dashboard`；支持 `7d/30d/90d`、地区筛选、学生/学习/内容/订单/激活码/反馈卡片、趋势、24h 活跃、题型分布、科目排行、地区统计、套餐销量和运营动态 |
 | 平台公共题库授权 | 可联调 | `/api/platform-admin/question-banks`、`question-bank-grants`；支持按 SaaS 套餐、指定租户或全部活跃租户披露平台公共题库 |
 | 租户采纳/同步公共题库 | 可联调 | `/api/tenant-content/public-question-banks`、`public-question-banks/adopt`、`public-question-banks/sync`、`public-question-banks/conflicts`、`public-question-banks/conflicts/resolve`、`public-question-banks/conflicts/resolve-batch`、`tenant-content/notifications`；租户只能看到自己订阅/授权范围内题库，采纳后生成租户自己的题库、入口、集合和题目快照，可直接进入练习；平台更新后可手动或由 worker 自动同步，新增/更新和冲突会生成租户内容通知；租户自改题目会标记冲突并跳过；后台可查询最近一次冲突明细，并可单条或批量选择“采纳平台版本”/“保留本地版本”，操作会重新校验授权并写入逐条审计，冲突全部处理后相关通知自动 resolved |
-| 题库导出基础 | 可联调 | `/api/tenant-content/exports/questions`、`/api/tenant-content/exports/jobs`；支持按题目集合、内容入口或分类节点导出 JSON/试卷 payload，后端校验租户内容编辑权限、跨租户隔离、答案/解析开关、复合题子题脱敏、导出 job 和审计；PDF/Word 二进制与水印 worker 后续补 |
+| 题库导出 | 可联调 | `/api/tenant-content/exports/questions`、`/api/tenant-content/exports/jobs`；支持按题目集合、内容入口或分类节点导出 JSON/试卷 payload，也支持 `pdf/docx` 异步导出；后端校验租户内容编辑权限、跨租户隔离、答案/解析开关、复合题子题脱敏、导出 job 和审计；`apps/worker --job exports` 负责 PDF/Word 渲染、水印、生成 `content_assets`、记录 hash/size/assetId，前端通过资源签名接口下载/预览 |
 
 ## 销售、代理、CRM
 
@@ -168,7 +168,7 @@
 | 视频 JSON preview/import | 可联调 | 支持 `videos/items`，后端校验题目、科目、资源引用，导入后写入 `question_videos` |
 | Excel/CSV 导入 | 可联调 | 题目、单词、知识手册、分数线、视频已支持 CSV 和 `.xlsx` 解析，解析后复用 `content_import_jobs/items/issues` 管线并保留 `parser_metadata`；模板下载、字段映射 API、字段映射覆盖白名单、导入任务详情、异步 worker 状态、导入后复检已接入；Taro 租户内容页已接上传/粘贴 preview/import、字段别名编辑、异步任务轮询和复检详情第一版 |
 | 大批量异步导入 | 可联调 | `executionMode=async` 会将 preview job 置为 `pending`；`apps/worker --job imports` 抢占 queued job，复用 API 导入 executor，支持重试、清锁和审计 |
-| 题库导出任务 | 可联调 | `content_export_jobs` 记录导出范围、格式、题量、输出 hash、选项和执行人；当前返回 inline base64 JSON 文件，前端可先下载 `.json` 或交给后续 PDF/Word worker 渲染 |
+| 题库导出任务 | 可联调 | `content_export_jobs` 记录导出范围、格式、题量、输出 hash、选项、执行人、异步状态、重试次数和 `asset_id`；JSON 类导出返回 inline base64，`pdf/docx` 由 exports worker 渲染为资源台账文件 |
 | 公共题库自动同步增强 | 部分覆盖 | `apps/worker --job public-banks` 已可抢占待同步采纳记录、自动同步平台新增/更新题目、记录失败和审计；同步新增/更新和冲突会写入 `tenant_content_notifications`，租户后台已有通知列表、已读/忽略、单条和批量冲突处理第一版；后续需接入生产定时调度、失败告警和更完整运营消息 |
 
 ## 当前验证
@@ -189,6 +189,7 @@ npm run test:worker:imports
 npm run test:worker:crm
 npm run test:worker:commerce
 npm run test:worker:assets
+npm run test:worker:exports
 npm run test:worker:public-banks
 git diff --check
 ```
diff --git a/docs/refactor/backend-handoff-roadmap.md b/docs/refactor/backend-handoff-roadmap.md
index 528238ca..34f78b7c 100644
--- a/docs/refactor/backend-handoff-roadmap.md
+++ b/docs/refactor/backend-handoff-roadmap.md
@@ -23,12 +23,12 @@
 | 多租户底座 | 可联调 | 租户、域名、品牌、设置、RLS 基础、审计、Supabase JWT/API 身份映射 | 真实云端 Auth/JWKS 回归、生产 RLS 深测 |
 | 平台后台 | 基础完成 | 租户、套餐、订阅、账单、服务费、用量、公共题库授权、公共题库自动同步 worker、公共题库冲突单条/批量处理 API、公共题库同步通知第一版 | 自动计费、平台审计、更完整运营消息 |
 | 租户后台 | 可联调 | 品牌、域名、支付账户、登录配置、密钥掩码、活动、兑换码、优惠券、勋章管理/发放、成员权限、角色模板、菜单/模块/字段权限配置 API、班级/教师/学生范围权限；Taro 工作台已接权限驱动模块入口，学生运营页已接学生创建/更新、禁用/恢复、批量导入、批量分班、备注和跟进任务第一版，租户设置页已接角色模板和成员绑定操作台第一版，营销中心已接 CRM 配置/队列和分佣结算操作台第一版 | 更细的数据范围组合、成员批量运营、真实打款/导出/凭证和完整权限菜单 |
-| 题库与练习 | 可联调 | 内容入口、任意深度分类、题目集合、顺序/随机/全真模拟蓝图、组卷快照、客观题后端判分、主观题 `selfJudgedCorrect` 自评、阅读理解/案例分析 `subAnswers` 多小题判分、答题、错题、收藏、模考报告、排行榜、公共题库采纳快照、手动同步、自动同步 worker、冲突查询/单条和批量处理 API、公共题库同步通知、JSON/试卷 payload 导出 | 长题干/公式图片混排体验、PDF/Word 导出 worker、排行榜防刷/预聚合 |
+| 题库与练习 | 可联调 | 内容入口、任意深度分类、题目集合、顺序/随机/全真模拟蓝图、组卷快照、客观题后端判分、主观题 `selfJudgedCorrect` 自评、阅读理解/案例分析 `subAnswers` 多小题判分、答题、错题、收藏、模考报告、排行榜、公共题库采纳快照、手动同步、自动同步 worker、冲突查询/单条和批量处理 API、公共题库同步通知、JSON/试卷 payload 导出、PDF/Word 异步导出 worker、水印和资料发布路径 | 长题干/公式图片混排体验、导出模板精排、排行榜防刷/预聚合 |
 | 背单词 | 可联调 | 单元、单词、进度、收藏、统计、每日计划、JSON/CSV/Excel 导入、排行榜 | 更细复习参数 |
 | 知识手册 | 可联调 | 科目、章节、条目、Markdown 内容、嵌套 JSON/CSV/Excel 导入 | 富文本资源、版本管理、附件/PDF 关联 |
 | 分数线 | 可联调 | 院校、专业、动态字段、记录、年份、趋势、后台维护、JSON/CSV/Excel 导入 | 复杂筛选、AI 择校上下文 |
 | 视频解析 | 可联调 | 单题视频、批量查询、后台视频绑定、JSON/CSV/Excel 导入、会员播放权限、播放次数扣减、签名 URL 和播放日志 | 深度防盗链、动态水印、播放统计 |
-| 资料下载 | 部分完成 | 资源台账、SVIP 权限校验、`local_dev`/阿里云 OSS/腾讯 COS/Supabase Storage 上传下载签名、上传确认、PDF/图片预览签名、assets worker 复检异常下架 | PDF 渲染、CDN 防盗链、杀毒扫描、视频水印 |
+| 资料下载 | 部分完成 | 资源台账、SVIP 权限校验、`local_dev`/阿里云 OSS/腾讯 COS/Supabase Storage 上传下载签名、上传确认、PDF/图片预览签名、assets worker 复检异常下架、题库导出 PDF/Word 可生成 `content_assets` 并走签名下载/预览 | CDN 防盗链、杀毒扫描、视频水印、资料前端操作体验 |
 | 会员与订单 | 可联调 | 下单、订单详情/状态轮询、优惠券领取/抵扣、零元订单自动开通、手工确认权限保护、激活码预检查/兑换、微信支付、支付宝、微信/支付宝发起退款、微信/支付宝退款查询确认、微信/支付宝退款通知 webhook、支付/退款补偿 worker、权益发放 | 完整资金流水对账、异常订单运营台 |
 | 登录认证 | 可联调 | 短信 mock、阿里云/腾讯云短信 adapter、迁移期 session、Supabase Auth JWT、微信小程序登录、微信网页登录、QQ 登录、手机号绑定/换绑、OAuth 配置表 | 真实生产账号和回调域名联调 |
 | 销售/代理/CRM | 可联调 | 邀请码、首绑保护、团队关系、销售统计、CRM 配置/队列、钉钉/飞书/企微 worker、分佣规则、成员分佣比例、订单/激活码归因、结算生成、审核和打款状态；Taro 营销中心已接第一版操作台 | 小程序码真实生成、CRM 分配策略、结算导出、真实打款、凭证、财务复核和销售转化看板 |
@@ -76,7 +76,7 @@
 ### P0：上云测试和前端主链路前必须处理
 
 - 生产鉴权：API 已支持 Supabase Auth JWT；继续做真实云端 Auth/JWKS 回归、RLS 深测，并在生产关闭 `x-user-id` 与 `x-platform-admin-key` 兼容入口。
-- 对象存储：上传/下载签名已接入阿里云 OSS、腾讯云 COS、Supabase Storage；上传确认、PDF/图片预览签名和 assets worker 复检已完成，继续补 PDF 渲染、视频播放防盗链、杀毒扫描和水印。
+- 对象存储：上传/下载签名已接入阿里云 OSS、腾讯云 COS、Supabase Storage；上传确认、PDF/图片预览签名、assets worker 复检和题库导出 PDF/Word worker 已完成，继续补视频播放防盗链、杀毒扫描和动态水印。
 - 真实数据 dry-run：导出 PocketBase 用户、题库、单词、知识手册、分数线、订单、权益，先跑 `npm run pb:import:dry-run`，再跑迁移和校验报告。
 - 生产环境配置：`.env.example` 和 `npm run readiness:production` / `npm run readiness:production:db` 已补；继续补数据库迁移流程、备份恢复、日志、告警和 API 容器部署说明。
 - Taro scaffold：`apps/taro` 地基已建立；学生端、租户后台、平台后台第一批 H5 页面已接真实 API，学生端已接地区选择、错题/收藏复习、阅读理解/案例分析多小题作答、题目反馈、视频解析、练习/模考报告、收银台、订单详情和售后入口第一版，平台后台关键写操作第一版已接入，租户工作台已接权限驱动模块入口，租户学生运营页已接学生创建/更新、禁用/恢复、批量导入、批量分班、备注和跟进任务第一版，租户内容页已接公共题库采纳/同步、冲突查看、单条/批量采纳平台或保留本地、导入问题、字段模板预览/下载、上传/粘贴预览、字段别名覆盖、同步/异步导入、异步轮询和复检详情第一版，租户设置页已接角色模板创建/编辑/停用、成员绑定模板和权限可见性配置第一版，租户营销中心已接 CRM 配置/队列和分佣结算操作台第一版；下一步补公式图片混排、更细数据范围 UI、平台后台审计增强和小程序兼容验证。
diff --git a/docs/refactor/frontend-handoff-index.md b/docs/refactor/frontend-handoff-index.md
index 174ad99b..d5381a74 100644
--- a/docs/refactor/frontend-handoff-index.md
+++ b/docs/refactor/frontend-handoff-index.md
@@ -87,11 +87,11 @@
 | 工作台 | `apps/taro/src/pages/tenant-admin/workbench/index.tsx` | `tenant-admin/overview`、`tenant-admin/dashboard`、`tenant-admin/permissions` |
 | 数据看板 | `apps/taro/src/pages/tenant-admin/dashboard/index.tsx` | `tenant-admin/dashboard` |
 | 学生运营 | `apps/taro/src/pages/tenant-admin/students/index.tsx` | `tenant-admin/classes`、`tenant-admin/teachers`、`tenant-admin/students`、`students/bulk-upsert`、`students/status`、`classes/members/bulk-assign`、`students/notes`、`students/followups` |
-| 题库内容 | `apps/taro/src/pages/tenant-admin/content/index.tsx` | `tenant-content/content-entries`、`tenant-content/imports`、`imports/detail`、`imports/issues`、`imports/field-mapping`、`imports/templates`、`imports/post-check`、`tenant-content/public-question-banks`、`public-question-banks/adopt`、`public-question-banks/sync`、`public-question-banks/conflicts`、`public-question-banks/conflicts/resolve`、`public-question-banks/conflicts/resolve-batch`、`tenant-content/notifications`、`tenant-content/notifications/status` |
+| 题库内容 | `apps/taro/src/pages/tenant-admin/content/index.tsx` | `tenant-content/content-entries`、`tenant-content/imports`、`imports/detail`、`imports/issues`、`imports/field-mapping`、`imports/templates`、`imports/post-check`、`tenant-content/exports/questions`、`tenant-content/exports/jobs`、`tenant-content/assets/sign-download`、`tenant-content/assets/sign-preview`、`tenant-content/public-question-banks`、`public-question-banks/adopt`、`public-question-banks/sync`、`public-question-banks/conflicts`、`public-question-banks/conflicts/resolve`、`public-question-banks/conflicts/resolve-batch`、`tenant-content/notifications`、`tenant-content/notifications/status` |
 | 营销中心 | `apps/taro/src/pages/tenant-admin/marketing/index.tsx` | `tenant-admin/coupons`、`code-batches`、`activation-codes`、`crm/config`、`crm/queue`、`commission/settings`、`member-rate`、`summary`、`orders`、`settlements`、`settlements/generate`、`settlements/status` |
 | 租户设置 | `apps/taro/src/pages/tenant-admin/settings/index.tsx` | `tenant-admin/overview`、`domains`、`payment-accounts`、`auth-providers`、`permissions`、`GET/PUT role-templates`、`POST role-templates/disable`、`GET/PUT members`、`POST members/disable` |
 
-当前租户后台已有第一批运营操作：工作台按权限矩阵隐藏不可见模块；学生运营页支持学生创建/更新、状态禁用/恢复、批量导入、批量分班、学生备注、跟进任务和完成跟进；题库内容页支持公共题库采纳/同步、同步通知查看与已读/忽略、同步冲突查看、单条/批量采纳平台版本或保留本地版本、导入任务详情、异步 job 轮询、导入问题查看、字段映射/模板预览/下载、JSON/CSV/Excel 导入预览和执行、字段别名覆盖和导入后复检详情；营销中心支持 CRM 配置保存、队列按状态查看、分佣规则、成员分佣比例、分佣订单明细、结算单生成、审核通过/驳回和标记线下打款；租户设置页支持角色模板创建、编辑、停用、权限点、菜单、模块、字段、基础数据范围配置、成员搜索/新建、成员绑定模板、成员状态和额外权限覆盖。下一批需要继续补更精细的学生导入模板体验、更细数据范围 UI、真实打款 provider、结算导出和凭证。
+当前租户后台已有第一批运营操作：工作台按权限矩阵隐藏不可见模块；学生运营页支持学生创建/更新、状态禁用/恢复、批量导入、批量分班、学生备注、跟进任务和完成跟进；题库内容页支持公共题库采纳/同步、同步通知查看与已读/忽略、同步冲突查看、单条/批量采纳平台版本或保留本地版本、导入任务详情、异步 job 轮询、导入问题查看、字段映射/模板预览/下载、JSON/CSV/Excel 导入预览和执行、字段别名覆盖和导入后复检详情，也可接 JSON/试卷 payload 同步导出与 PDF/Word 异步导出 job 轮询，完成后用 `assetId` 走后台资源签名下载/预览；营销中心支持 CRM 配置保存、队列按状态查看、分佣规则、成员分佣比例、分佣订单明细、结算单生成、审核通过/驳回和标记线下打款；租户设置页支持角色模板创建、编辑、停用、权限点、菜单、模块、字段、基础数据范围配置、成员搜索/新建、成员绑定模板、成员状态和额外权限覆盖。下一批需要继续补更精细的学生导入模板体验、更细数据范围 UI、真实打款 provider、结算导出和凭证。
 
 ## 已落地的 Taro 平台后台页面
 
diff --git a/docs/refactor/implementation-status.md b/docs/refactor/implementation-status.md
index edabbe07..d02dddb0 100644
--- a/docs/refactor/implementation-status.md
+++ b/docs/refactor/implementation-status.md
@@ -4,7 +4,7 @@
 
 ## 当前结论
 
-当前重构已经完成了 Supabase/PostgreSQL 多租户底座、核心业务表、PocketBase 数据导入器雏形、学生端核心 API、租户后台 API、平台后台 SaaS 账务 API、内容资产/题目/单词/知识手册/分数线/视频 JSON/CSV/Excel 批量导入基础闭环、异步导入 worker、题库导出基础、题库入口/任意深度分类/题目集合/练习蓝图/组卷快照基础闭环，以及本地 Docker/API 构建验证。
+当前重构已经完成了 Supabase/PostgreSQL 多租户底座、核心业务表、PocketBase 数据导入器雏形、学生端核心 API、租户后台 API、平台后台 SaaS 账务 API、内容资产/题目/单词/知识手册/分数线/视频 JSON/CSV/Excel 批量导入基础闭环、异步导入 worker、题库导出基础、PDF/Word 导出 worker、题库入口/任意深度分类/题目集合/练习蓝图/组卷快照基础闭环，以及本地 Docker/API 构建验证。
 
 但这还不是完整商用交付状态，也不能说旧项目核心功能已经全部重构完成。现在更准确的状态是：后端商用架构骨架已经立住，核心业务正在按模块补齐。部分功能已经有可调用 API，部分功能只有数据模型和导入映射，部分功能还没有前端/自动化测试闭环。
 
@@ -24,7 +24,7 @@
 | 模块 | 数据模型 | PocketBase 导入 | API | 自动化测试 | 当前状态 |
 | --- | --- | --- | --- | --- | --- |
 | 多租户隔离 | 已建 `tenants`、`tenant_domains`、`tenant_branding`、`tenant_settings`、RLS 基础 | 部分支持 | 租户解析、品牌、域名、支付账户、登录 provider、平台建租户已实现 | 核心 API 集成测试含租户隔离断言 | 基础可用，正式 JWT/RLS 权限闭环未完成 |
-| 刷题题库 | 已建题库、题目、题目版本、内容入口、任意深度分类树、考试意向标记、题目集合、练习蓝图、导入任务台账、导出任务台账、公共题库授权/采纳表、租户内容通知表 | 已支持核心映射，JSON/CSV/Excel 导入可落到新入口/节点/集合，阅读理解/案例分析子题沿用 `subQuestions/sub_questions` | 题目列表、内容入口、分类树、集合题目、顺序/随机/全真模拟 session、答题提交、复合题 `subAnswers` 判分和报告明细、租户后台题目录入/更新、JSON/CSV/Excel 预览/导入、JSON/试卷 payload 导出、异步导入 worker、平台公共题库授权、租户采纳快照、手动同步、自动同步 worker、同步通知、冲突查询和单条/批量冲突处理已实现 | 核心 API 集成测试含导航、组卷、复合题后台录入/练习/判分/报告、导入、导出权限/脱敏、公共题库授权、采纳后组卷、同步新增题、通知隔离/已读/自动 resolved、租户自改冲突保护、单条/批量冲突处理和 worker 自动同步断言 | 新题库导航和组卷基础闭环可跑，阅读理解/案例分析多小题第一版可联调，公共题库采纳/手动/自动同步、同步通知、冲突查询/处理、导入后复检、模板下载、字段映射 API 和导出基础可联调；PDF/Word 导出 worker、公共题库生产调度/失败告警和更完整运营消息仍需补齐 |
+| 刷题题库 | 已建题库、题目、题目版本、内容入口、任意深度分类树、考试意向标记、题目集合、练习蓝图、导入任务台账、导出任务台账、公共题库授权/采纳表、租户内容通知表 | 已支持核心映射，JSON/CSV/Excel 导入可落到新入口/节点/集合，阅读理解/案例分析子题沿用 `subQuestions/sub_questions` | 题目列表、内容入口、分类树、集合题目、顺序/随机/全真模拟 session、答题提交、复合题 `subAnswers` 判分和报告明细、租户后台题目录入/更新、JSON/CSV/Excel 预览/导入、JSON/试卷 payload 导出、PDF/Word 异步导出 worker、异步导入 worker、平台公共题库授权、租户采纳快照、手动同步、自动同步 worker、同步通知、冲突查询和单条/批量冲突处理已实现 | 核心 API 集成测试含导航、组卷、复合题后台录入/练习/判分/报告、导入、导出权限/脱敏、异步 PDF/Word job 创建、exports worker、公共题库授权、采纳后组卷、同步新增题、通知隔离/已读/自动 resolved、租户自改冲突保护、单条/批量冲突处理和 worker 自动同步断言 | 新题库导航和组卷基础闭环可跑，阅读理解/案例分析多小题第一版可联调，公共题库采纳/手动/自动同步、同步通知、冲突查询/处理、导入后复检、模板下载、字段映射 API、JSON/PDF/Word 导出可联调；公共题库生产调度/失败告警、更精细导出模板和更完整运营消息仍需补齐 |
 | 错题本 | 已建 `wrong_questions` | 已支持旧错题归一化 | 错题列表、答题自动入错题、移出错题已实现 | 仅烟测 | 基础功能已实现，复习计划和统计未完成 |
 | 收藏夹 | 已建 `favorite_questions` | 已支持旧收藏归一化 | 收藏/取消收藏、收藏列表已实现 | 仅烟测 | 基础功能已实现 |
 | 用户订阅/题库会员/SVIP | 已建 `orders`、`payments`、`entitlements`、`svip_plans`、激活码 | 已映射旧 SVIP/会员权益 | 下单、订单详情/状态轮询、手工支付确认权限保护、微信/支付宝支付、微信/支付宝发起退款、微信/支付宝退款查询确认、微信/支付宝退款通知 webhook、激活码预检查/兑换、优惠券抵扣、零元订单自动开通、权益查询已实现 | API 集成测试 | 商城主链路可联调，对账、支付补偿和异常订单自动处理待补 |
@@ -32,7 +32,7 @@
 | 知识手册 | 已建手册科目、章节、条目，并可绑定 `content_entries/content_nodes` | 已支持内容导入 | 只读 API、租户后台手册科目/章节/条目维护 API、嵌套 JSON 预览导入已实现 | 核心 API 集成测试含导入断言 | 学生端阅读、后台维护和批量 JSON 导入基础可用，富文本资源/版本管理待补 |
 | 分数线 | 已建院校、专业、字段、记录表 | 已支持导入映射 | 字段、院校、专业、记录、趋势、年份、租户后台维护 API、JSON 预览导入已实现 | 核心 API 集成测试含导入断言 | 查询、后台维护和批量 JSON 导入基础闭环已实现，复杂动态筛选和 AI 择校上下文待补 |
 | 题目视频讲解 | 已建 `video_explanations`、`question_videos` | 已支持导入映射 | 单题视频、批量预加载、通用视频搜索、播放签名、视频次数扣减、租户后台视频创建绑定 API、JSON 预览导入已实现 | 核心 API 集成测试含播放和导入断言 | 播放、权益、后台绑定和批量 JSON 导入链路已实现，深度防盗链、动态水印和播放统计待补 |
-| 资料下载/PDF | 已扩展 `content_assets`，新增资源台账和导入任务表 | 旧 `app_assets/images` 兼容导入 | 租户后台资源管理、OSS/COS/Supabase Storage 上传/下载签名、上传确认、PDF/图片预览签名、学生端资料列表/下载权限已实现 | 核心 API 集成测试含 SVIP 资料下载，assets worker 测试 | 资料资源基础闭环可跑，CDN 防盗链、杀毒扫描和资料下载前端待补 |
+| 资料下载/PDF | 已扩展 `content_assets`，新增资源台账和导入任务表 | 旧 `app_assets/images` 兼容导入 | 租户后台资源管理、OSS/COS/Supabase Storage 上传/下载签名、上传确认、PDF/图片预览签名、题库导出 PDF/Word 自动发布资源、学生端资料列表/下载权限已实现 | 核心 API 集成测试含 SVIP 资料下载，assets worker 和 exports worker 测试 | 资料资源基础闭环可跑，CDN 防盗链、杀毒扫描和资料下载前端待补 |
 | 个人中心 | 已建 `student_profiles`、会员权益、订单、练习记录、`badges/user_badges` | 已支持部分用户资料和勋章导入 | 个人资料、目标院校/专业、手机号绑定/换绑、会员状态、最近练习、统计聚合、签到积分、题目反馈、考试倒计时、勋章 API 已实现 | API 集成测试 | 学生端基础个人中心已实现，学习报告可视化和更细任务系统待补 |
 | 活动/优惠 | 已建优惠券、激活码、激活码批次、banner、FAQ、公告、勋章等基础表 | 部分支持 | banner/FAQ/公告只读与租户后台维护、激活码预检查/兑换、激活码批次、批量生成激活码、优惠券维护、前台领取/下单抵扣、勋章维护和手动发放已实现 | 核心 API 集成测试 | 基础运营后台可用，勋章自动发放、复杂活动规则、营销自动化、核销报表待补 |
 | 销售/代理客资追踪 | 已建推荐码、首绑客资、团队关系、小程序码缓存、CRM 队列 | 旧 `referral_tracks` 已有映射基础 | 邀请码、扫码/分享事件、首绑保护、销售统计、客资明细、手动补绑、团队关系、CRM 配置/队列、CRM worker 推送已实现 | 核心 API 集成测试、CRM worker 集成测试 | 增长链路基础可用，真实微信小程序码、CRM 分配策略、富卡片和销售转化看板待补 |
diff --git a/docs/refactor/legacy-feature-gap-matrix.md b/docs/refactor/legacy-feature-gap-matrix.md
index 3bc8ff12..3380e022 100644
--- a/docs/refactor/legacy-feature-gap-matrix.md
+++ b/docs/refactor/legacy-feature-gap-matrix.md
@@ -59,7 +59,7 @@
 | 激活码 | 已覆盖 | 批次、生成、预检查、兑换、自用码拒绝、地区校验主链路已有 |
 | 勋章管理 | 部分覆盖 | 后台勋章维护、手动发放、重复发放幂等、学生端勋章展示和权限隔离已覆盖；自动发放规则、积分活动联动和前端运营 UI 待补 |
 | 题库录入 | 已覆盖 | 单题创建/更新、题目/单词/知识手册/分数线/视频 JSON/CSV/Excel 同步/异步导入、集合/蓝图、导入后复检、模板下载、字段映射 API 和导入任务详情已有；Taro 租户内容页已接上传/粘贴预览、模板下载、字段别名编辑、同步/异步执行、异步轮询和复检结果详情第一版；真实数据验收待补 |
-| 题库导出 PDF/Word/JSON | 部分覆盖 | 服务端 JSON、`paper_json` 和打印 payload 导出基础已补，含租户内容编辑权限、跨租户拒绝、答案/解析开关、子题脱敏、导出 job 和审计；PDF/Word 二进制、水印、发布到资料下载和导出 worker 待补 |
+| 题库导出 PDF/Word/JSON | 部分覆盖 | 服务端 JSON、`paper_json`、打印 payload、PDF 和 Word 异步导出已补，含租户内容编辑权限、跨租户拒绝、答案/解析开关、子题脱敏、导出 job、审计、水印、`content_assets` 发布/下载路径和 exports worker；后续补更精细试卷模板、每日一练素材导出和后台操作台体验 |
 | 题型分组/模拟卷配置 | 部分覆盖 | question_type_groups 表和 blueprint 有基础；后台配置体验待补 |
 | 背单词维护 | 已覆盖 | 单元/单词 CRUD 和导入已有 |
 | 知识手册维护 | 已覆盖 | subject/chapter/entry CRUD 和导入已有 |
@@ -98,7 +98,7 @@
 1. 阅读理解/案例分析多小题：后端 `subAnswers` 判分、断点恢复、报告 `subResults` 和 Taro 作答 UI 第一版已完成；仍需优化长题干排版、公式图片混排和复盘体验。
 2. 排行榜增强：刷题、模考、背单词、积分排行榜主接口已有；还需防刷、日/周榜预聚合、运营后台排名看板。
 3. 账号设置完整流：绑定/更换手机号基础 API 已完成；仍缺头像上传、微信/QQ 账号合并、密码/邮箱能力。
-4. 题库导出：服务端 JSON/试卷 payload 导出、权限审计和答案脱敏已补；仍缺 PDF/Word 二进制生成、水印、资料发布和后台导出操作台。
+4. 题库导出：服务端 JSON/试卷 payload、PDF/Word 二进制生成、水印、资料发布路径、权限审计和答案脱敏已补；仍缺更精细试卷模板、每日一练素材导出和后台导出操作台体验。
 5. 导入扩展：题目/单词/知识手册/分数线/视频已支持 JSON、CSV 和 Excel 预览导入，并可用 `executionMode=async` 进入 imports worker；导入后复检、导入任务详情、模板下载按钮、字段映射 API、Taro 字段别名编辑、异步轮询和 PocketBase JSON dry-run 报告已补，仍缺真实数据执行验收。
 6. 公共题库商业化：平台公共/地区题库授权、租户快照采纳、手动同步、自动同步 worker、同步通知、冲突查询、租户自改冲突保护和单条/批量冲突处理已完成基础闭环；还需生产定时调度、失败告警和更完整运营后台消息。
 7. CRM/销售结算：CRM worker、分佣规则、结算单、审核和打款状态基础闭环已完成；仍缺轮询/定向分配、打款导出、凭证和销售结算看板。
@@ -127,4 +127,4 @@
 1. 主题模板、素材库、主题预览/发布。
 2. 班级、教师、学生分组和学习督导。
 3. AI 择校推荐和 PDF 报告。
-4. 题库导出二进制生成、试卷排版 worker、每日一练运营工具。
+4. 题库导出模板精排、每日一练运营工具。
diff --git a/docs/refactor/next-development-todo.md b/docs/refactor/next-development-todo.md
index 9e4771b5..449bb94e 100644
--- a/docs/refactor/next-development-todo.md
+++ b/docs/refactor/next-development-todo.md
@@ -27,6 +27,7 @@
 - 租户后台数据看板已完成首版聚合 API：`GET /api/tenant-admin/dashboard`，支持租户/地区维度的收益、注册、学习、内容、激活码、反馈、趋势、24h 活跃、套餐销量和运营动态，前端可直接联调。
 - 支付/退款补偿 worker 已完成：`apps/worker --job commerce` 可查询微信/支付宝支付和处理中退款，补偿漏通知订单，支付成功幂等开通权益，退款成功幂等更新退款/订单/支付并在全额退款时撤销订单权益。
 - 内容资源复检 worker 已完成：`apps/worker --job assets` 可复检 `content_assets` 中的托管对象元数据，正常资源写回复检证据，异常资源自动置为 `failed + draft` 并写入审计和安全标记。
+- 题库导出 worker 已完成：`apps/worker --job exports` 可抢占 `pdf/docx` 导出任务，渲染 PDF/Word、水印、写入对象存储或本地开发存储，创建 `content_assets` 并回填 `assetId/hash/size`。
 - 本地验证：`npm run check:refactor` 已通过。
 
 当前更适合进入前端联调前阅读的总览文档：
@@ -83,8 +84,8 @@
 
 4. 题库导出增强
    - 已完成服务端 JSON、`paper_json`、`print_payload` 基础导出和导出 job 台账。
-   - 当前前端可先下载 inline base64 JSON，或把 `paper_json` 交给后台打印预览组件渲染。
-   - 继续补 PDF/Word 二进制生成 worker、页面水印、导出发布到资料下载、导出模板和操作台。
+   - 已完成 PDF/Word 二进制生成 worker、页面水印、导出发布到 `content_assets`、后台资源签名下载/预览路径。
+   - 继续补更精细的试卷模板、每日一练素材导出、导出操作台和多模板排版。
 
 5. 公共题库和租户授权
    - 已完成平台公共题库/地区题库的基础授权、租户采纳、题目快照复制和手动同步。
@@ -219,5 +220,5 @@
 3. 补平台后台增强：租户详情/编辑、平台审计报表、自动计费、账单批量操作和更细平台权限点。
 4. 云服务器部署 Supabase/PostgreSQL 和 API，配置对象存储生产环境变量，跑 `check:refactor` 的远程等价测试。
 5. 导出现有 PocketBase 数据，做完整 dry-run 迁移。
-6. 并行补对象存储、真实登录、完整资金流水对账、题库导出 PDF/Word worker、公共题库生产定时调度和失败告警。
+6. 并行补对象存储、真实登录、完整资金流水对账、题库导出模板精排/每日一练、公共题库生产定时调度和失败告警。
 7. 前后端联调通过后，再做支付、权限、数据导入、资料下载、视频播放的商用验收。
diff --git a/docs/refactor/object-storage.md b/docs/refactor/object-storage.md
index d0dfd089..9fb6d49c 100644
--- a/docs/refactor/object-storage.md
+++ b/docs/refactor/object-storage.md
@@ -70,6 +70,12 @@ POST /api/tenant-content/assets/sign-preview
 npm --workspace @tiku-saas/worker run assets:once
 ```
 
+题库 PDF/Word 导出 worker：
+
+```bash
+npm --workspace @tiku-saas/worker run exports:once
+```
+
 ## 标准上传流程
 
 后台前端上传 PDF、图片、视频或资料包时必须走下面流程：
@@ -83,6 +89,19 @@ npm --workspace @tiku-saas/worker run assets:once
 7. 生产环境定时运行 assets worker，复检 `pending/verified` 托管对象的大小、MIME、SHA-256 等元数据。
 8. 如果复检发现对象丢失、跨租户 objectKey、大小/MIME/checksum 不一致，worker 会把资源置为 `uploadStatus=failed`，并将 `active` 资源退回 `draft`，同时写入 `security_flags.assetRecheckFailed=true` 和 `audit_logs`。
 
+## 题库导出生成资源
+
+题库导出 PDF/Word 不走前端上传，而是由后端和 worker 自动进入资源台账：
+
+1. 租户后台调用 `POST /api/tenant-content/exports/questions`，`format=pdf` 或 `format=docx`。
+2. API 校验租户内容权限、导出范围和答案/解析开关，创建 `content_export_jobs.status=pending`。
+3. `apps/worker --job exports` 抢占 pending job，复用后端导出 payload 规则渲染 PDF/Word，应用 `options.watermarkText` 水印。
+4. worker 将二进制文件写入配置的对象存储。`local_dev` 会写入 `EXPORT_LOCAL_STORAGE_ROOT`，生产建议使用 `supabase_storage`、`aliyun_oss` 或 `tencent_cos`。
+5. worker 创建 `content_assets`，设置 `uploadStatus=verified`，并把 `assetId`、文件名、大小和 SHA-256 回填到 `content_export_jobs.output_metadata`。
+6. 前端轮询 `GET /api/tenant-content/exports/jobs`，拿到 `assetId` 后通过 `POST /api/tenant-content/assets/sign-download` 或 `sign-preview` 下载/预览。
+
+注意：PDF/Word 文件本身不直接存入数据库，数据库只保存 job、资源台账和校验 metadata。
+
 确认上传示例：
 
 ```json
diff --git a/docs/refactor/taro-frontend-integration.md b/docs/refactor/taro-frontend-integration.md
index ab624671..5348ae60 100644
--- a/docs/refactor/taro-frontend-integration.md
+++ b/docs/refactor/taro-frontend-integration.md
@@ -1015,7 +1015,10 @@ failed：导入结果和目标表不一致，必须提示管理员排查，不
 
 ## 题库导出对接
 
-租户后台题库导出统一走后端生成结构化 payload，前端不要直接查 Supabase 表拼导出文件。当前后端已支持 JSON、`paper_json` 和 `print_payload` 三类基础导出，适合先做后台“导出 JSON/试卷预览”功能；PDF/Word 二进制、水印和发布到资料下载后续由 worker 增强。
+租户后台题库导出统一走后端生成结构化 payload，前端不要直接查 Supabase 表拼导出文件。当前后端支持两种交付形态：
+
+- 同步结构化导出：`json`、`paper_json`、`print_payload`，接口直接返回 base64 JSON/payload。
+- 异步二进制导出：`pdf`、`docx`，接口先返回 `pending` job，由 `apps/worker --job exports` 渲染 PDF/Word、水印并写入 `content_assets`，前端轮询 job 后再走资源签名下载或预览。
 
 可用接口：
 
@@ -1066,6 +1069,27 @@ GET  /api/tenant-content/exports/jobs
 }
 ```
 
+PDF/Word 异步导出：
+
+```json
+{
+  "scopeType": "collection",
+  "scopeId": "<questionCollectionId>",
+  "format": "pdf",
+  "exportType": "paper",
+  "includeAnswers": false,
+  "includeExplanations": false,
+  "options": {
+    "title": "天津专升本模拟试卷",
+    "watermarkText": "仅供内部使用",
+    "publishToAssets": true,
+    "assetVisibility": "tenant"
+  }
+}
+```
+
+`format` 可为 `pdf` 或 `docx`。`publishToAssets=true` 表示导出文件可作为资料资源展示给对应可见范围用户；不传时默认生成后台私有资源，仅后台可下载。`assetVisibility` 支持 `public`、`tenant`、`members`、`svip`、`private`，生产默认建议用 `tenant/private/svip`，不要轻易公开带题目的文件。
+
 响应关键结构：
 
 ```json
@@ -1102,14 +1126,54 @@ GET  /api/tenant-content/exports/jobs
 }
 ```
 
+异步二进制导出初始响应：
+
+```json
+{
+  "job": {
+    "id": "...",
+    "status": "pending",
+    "questionCount": 20,
+    "outputHash": "..."
+  },
+  "export": null
+}
+```
+
+worker 完成后，`GET /api/tenant-content/exports/jobs` 返回：
+
+```json
+{
+  "items": [
+    {
+      "id": "...",
+      "format": "pdf",
+      "status": "completed",
+      "assetId": "...",
+      "attemptCount": 1,
+      "outputMetadata": {
+        "delivery": "content_asset",
+        "fileName": "天津专升本模拟试卷.pdf",
+        "mimeType": "application/pdf",
+        "sizeBytes": 123456,
+        "checksumSha256": "...",
+        "assetId": "..."
+      }
+    }
+  ]
+}
+```
+
 前端处理规则：
 
 - 导出按钮只给具备租户内容编辑权限的后台成员展示；接口仍以后端 `TENANT_CONTENT_EDITOR_REQUIRED` 为准。
 - 下载 JSON 时使用 `files[0].contentBase64` 生成 Blob，文件名使用后端返回的 `filename`。
 - `includeAnswers=false` 时，顶层答案字段和阅读理解/案例分析的子题答案都会被后端脱敏；前端不要在本地重新合并答案。
 - `includeExplanations=false` 时，不展示解析，也不要从题目详情接口额外补解析。
-- `paper_json` 可先用于后台试卷预览和打印；正式 PDF/Word 导出等后端 worker 完成后再接二进制文件下载。
-- `GET /api/tenant-content/exports/jobs?scopeType=collection&scopeId=...` 用于后台导出历史；当前记录 metadata 和输出 hash，不长期保存完整导出内容。
+- `paper_json` 可用于后台试卷预览和打印；`pdf/docx` 用于正式文件下载、资料发布和带水印留档。
+- `GET /api/tenant-content/exports/jobs?scopeType=collection&scopeId=...` 用于后台导出历史；结构化导出只记录 metadata 和输出 hash，二进制导出记录 `assetId`、文件名、大小和 checksum。
+- `status=pending/rendering` 时展示生成中；`completed` 且存在 `assetId` 后，后台可调用 `POST /api/tenant-content/assets/sign-download` 下载，PDF 也可调用 `POST /api/tenant-content/assets/sign-preview` 预览。
+- 学生资料页只展示 `content_assets.status=active` 且非 `private` 的资源；后台私有导出不会出现在学生资料列表。
 - 跨租户导出会返回 404 或 403，前端不要重试其它租户 ID。
 
 ## 公共题库采纳对接
diff --git a/package-lock.json b/package-lock.json
index a013a48b..d6e4e132 100644
--- a/package-lock.json
+++ b/package-lock.json
@@ -86,10 +86,13 @@
       "dependencies": {
         "@supabase/storage-js": "^2.108.2",
         "ali-oss": "^6.23.0",
+        "docx": "^9.7.1",
+        "pdfkit": "^0.19.1",
         "pg": "^8.16.3"
       },
       "devDependencies": {
         "@types/node": "^24.0.4",
+        "@types/pdfkit": "^0.17.6",
         "@types/pg": "^8.15.4",
         "tsx": "^4.20.3",
         "typescript": "^5.8.3"
@@ -2895,6 +2898,30 @@
       "dev": true,
       "license": "MIT"
     },
+    "node_modules/@noble/ciphers": {
+      "version": "1.3.0",
+      "resolved": "https://registry.npmjs.org/@noble/ciphers/-/ciphers-1.3.0.tgz",
+      "integrity": "sha512-2I0gnIVPtfnMw9ee9h1dJG7tp81+8Ob3OJb3Mv37rx5L40/b0i7djjCVvGOVqc9AEIQyvyu1i6ypKdFw8R8gQw==",
+      "license": "MIT",
+      "engines": {
+        "node": "^14.21.3 || >=16"
+      },
+      "funding": {
+        "url": "https://paulmillr.com/funding/"
+      }
+    },
+    "node_modules/@noble/hashes": {
+      "version": "1.8.0",
+      "resolved": "https://registry.npmjs.org/@noble/hashes/-/hashes-1.8.0.tgz",
+      "integrity": "sha512-jCs9ldd7NwzpgXDIf6P3+NrHh9/sD6CQdxHyjQI+h/6rDNo88ypBxxz45UDuZHz9r3tNz7N/VInSVoVdtXEI4A==",
+      "license": "MIT",
+      "engines": {
+        "node": "^14.21.3 || >=16"
+      },
+      "funding": {
+        "url": "https://paulmillr.com/funding/"
+      }
+    },
     "node_modules/@nodelib/fs.scandir": {
       "version": "2.1.5",
       "resolved": "https://registry.npmjs.org/@nodelib/fs.scandir/-/fs.scandir-2.1.5.tgz",
@@ -3777,6 +3804,16 @@
       "dev": true,
       "license": "Apache-2.0"
     },
+    "node_modules/@swc/helpers": {
+      "version": "0.5.23",
+      "resolved": "https://registry.npmjs.org/@swc/helpers/-/helpers-0.5.23.tgz",
+      "integrity": "sha512-5lSsMOTXURePglDfvuAQUqkGek9Hg2kksOYay2m0+XR++b2NWYL/4sWyuvVBIs8oKnJaxkdi9whaL/sqN13afw==",
+      "license": "Apache-2.0",
+      "peer": true,
+      "dependencies": {
+        "tslib": "^2.8.0"
+      }
+    },
     "node_modules/@swc/register": {
       "version": "0.1.10",
       "resolved": "https://registry.npmjs.org/@swc/register/-/register-0.1.10.tgz",
@@ -4981,7 +5018,6 @@
       "version": "25.6.0",
       "resolved": "https://registry.npmjs.org/@types/node/-/node-25.6.0.tgz",
       "integrity": "sha512-+qIYRKdNYJwY3vRCZMdJbPLJAtGjQBudzZzdzwQYkEPQd+PJGixUL5QfvCLDaULoLv+RhT3LDkwEfKaAkgSmNQ==",
-      "dev": true,
       "license": "MIT",
       "peer": true,
       "dependencies": {
@@ -4998,6 +5034,16 @@
         "@types/node": "*"
       }
     },
+    "node_modules/@types/pdfkit": {
+      "version": "0.17.6",
+      "resolved": "https://registry.npmjs.org/@types/pdfkit/-/pdfkit-0.17.6.tgz",
+      "integrity": "sha512-tIwzxk2uWKp0Cq9JIluQXJid77lYhF52EsIOwhsMF4iWLA6YneoBR1xVKYYdAysHuepUB0OX4tdwMiUDdGKmig==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@types/node": "*"
+      }
+    },
     "node_modules/@types/pg": {
       "version": "8.20.0",
       "resolved": "https://registry.npmjs.org/@types/pg/-/pg-8.20.0.tgz",
@@ -6312,6 +6358,24 @@
         "node": ">=8"
       }
     },
+    "node_modules/brotli": {
+      "version": "1.3.3",
+      "resolved": "https://registry.npmjs.org/brotli/-/brotli-1.3.3.tgz",
+      "integrity": "sha512-oTKjJdShmDuGW94SyyaoQvAjf30dZaHnjJ8uAF+u2/vGJkJbJPJAT1gDiOJP5v1Zb6f9KEyW/1HpuaWIXtGHPg==",
+      "license": "MIT",
+      "dependencies": {
+        "base64-js": "^1.1.2"
+      }
+    },
+    "node_modules/browserify-zlib": {
+      "version": "0.2.0",
+      "resolved": "https://registry.npmjs.org/browserify-zlib/-/browserify-zlib-0.2.0.tgz",
+      "integrity": "sha512-Z942RysHXmJrhqk88FmKBVq/v5tqmSkDz7p54G/MGyjMnCFFnC79XWNbg+Vta8W6Wb2qtSZTSxIGkJrRpCFEiA==",
+      "license": "MIT",
+      "dependencies": {
+        "pako": "~1.0.5"
+      }
+    },
     "node_modules/browserslist": {
       "version": "4.28.4",
       "resolved": "https://registry.npmjs.org/browserslist/-/browserslist-4.28.4.tgz",
@@ -8211,6 +8275,12 @@
         "node": ">= 4.0.0"
       }
     },
+    "node_modules/dfa": {
+      "version": "1.2.0",
+      "resolved": "https://registry.npmjs.org/dfa/-/dfa-1.2.0.tgz",
+      "integrity": "sha512-ED3jP8saaweFTjeGX8HQPjeC1YYyZs98jGNZx6IiBvxW7JG5v492kamAQB3m2wop07CvU/RQmzcKr6bgcC5D/Q==",
+      "license": "MIT"
+    },
     "node_modules/digest-header": {
       "version": "1.1.0",
       "resolved": "https://registry.npmjs.org/digest-header/-/digest-header-1.1.0.tgz",
@@ -8256,6 +8326,41 @@
         "node": ">=6.0.0"
       }
     },
+    "node_modules/docx": {
+      "version": "9.7.1",
+      "resolved": "https://registry.npmjs.org/docx/-/docx-9.7.1.tgz",
+      "integrity": "sha512-ilXFf9Moz47ABjFpDiA5s1w9lpb4EFSp7+5iiJSbfyYDM+bpZdAgLlSr7fW4aXhVe/E+F6QCv0EvRVFEd5CsWg==",
+      "license": "MIT",
+      "dependencies": {
+        "@types/node": "^25.2.3",
+        "hash.js": "^1.1.7",
+        "jszip": "^3.10.1",
+        "nanoid": "^5.1.3",
+        "xml": "^1.0.1",
+        "xml-js": "^1.6.8"
+      },
+      "engines": {
+        "node": ">=10"
+      }
+    },
+    "node_modules/docx/node_modules/nanoid": {
+      "version": "5.1.16",
+      "resolved": "https://registry.npmjs.org/nanoid/-/nanoid-5.1.16.tgz",
+      "integrity": "sha512-kVrnsrJqMR8+oLJnGEmSWw9BivK5mt7H3FZatVRjrc5wGqFYuBxX1yG7+A7Gi5AefkX6t/oCkizcQgpu0cY1dQ==",
+      "funding": [
+        {
+          "type": "github",
+          "url": "https://github.com/sponsors/ai"
+        }
+      ],
+      "license": "MIT",
+      "bin": {
+        "nanoid": "bin/nanoid.js"
+      },
+      "engines": {
+        "node": "^18 || >=20"
+      }
+    },
     "node_modules/dom-converter": {
       "version": "0.2.0",
       "resolved": "https://registry.npmjs.org/dom-converter/-/dom-converter-0.2.0.tgz",
@@ -9616,7 +9721,6 @@
       "version": "3.1.3",
       "resolved": "https://registry.npmjs.org/fast-deep-equal/-/fast-deep-equal-3.1.3.tgz",
       "integrity": "sha512-f3qQ9oQy9j2AhBe/H9VC91wLmKBCCU/gDOnKNAYG5hswO7BLKj09Hc5HYNz9cGI++xlpDCIgDaitVs03ATR84Q==",
-      "dev": true,
       "license": "MIT"
     },
     "node_modules/fast-glob": {
@@ -9958,6 +10062,32 @@
         }
       }
     },
+    "node_modules/fontkit": {
+      "version": "2.0.4",
+      "resolved": "https://registry.npmjs.org/fontkit/-/fontkit-2.0.4.tgz",
+      "integrity": "sha512-syetQadaUEDNdxdugga9CpEYVaQIxOwk7GlwZWWZ19//qW4zE5bknOKeMBDYAASwnpaSHKJITRLMF9m1fp3s6g==",
+      "license": "MIT",
+      "dependencies": {
+        "@swc/helpers": "^0.5.12",
+        "brotli": "^1.3.2",
+        "clone": "^2.1.2",
+        "dfa": "^1.2.0",
+        "fast-deep-equal": "^3.1.3",
+        "restructure": "^3.0.0",
+        "tiny-inflate": "^1.0.3",
+        "unicode-properties": "^1.4.0",
+        "unicode-trie": "^2.0.0"
+      }
+    },
+    "node_modules/fontkit/node_modules/clone": {
+      "version": "2.1.2",
+      "resolved": "https://registry.npmjs.org/clone/-/clone-2.1.2.tgz",
+      "integrity": "sha512-3Pe/CF1Nn94hyhIYpjtiLhdCoEoz0DqQ+988E9gmeEdQZlojxnOb74wctFyuwWQHzqyf9X7C7MG8juUpqBJT8w==",
+      "license": "MIT",
+      "engines": {
+        "node": ">=0.8"
+      }
+    },
     "node_modules/for-each": {
       "version": "0.3.5",
       "resolved": "https://registry.npmjs.org/for-each/-/for-each-0.3.5.tgz",
@@ -10531,6 +10661,16 @@
         "url": "https://github.com/sponsors/ljharb"
       }
     },
+    "node_modules/hash.js": {
+      "version": "1.1.7",
+      "resolved": "https://registry.npmjs.org/hash.js/-/hash.js-1.1.7.tgz",
+      "integrity": "sha512-taOaskGt4z4SOANNseOviYDvjEJinIkRgmp7LbKP2YTTmVxWBl87s/uzK9r+44BclBSp2X7K1hqeNfz9JbBeXA==",
+      "license": "MIT",
+      "dependencies": {
+        "inherits": "^2.0.3",
+        "minimalistic-assert": "^1.0.1"
+      }
+    },
     "node_modules/hasown": {
       "version": "2.0.4",
       "resolved": "https://registry.npmjs.org/hasown/-/hasown-2.0.4.tgz",
@@ -11663,6 +11803,12 @@
       "integrity": "sha512-pZe//GGmwJndub7ZghVHz7vjb2LgC1m8B07Au3eYqeqv9emhESByMXxaEgkUkEqJe87oBbSniGYoQNIBklc7IQ==",
       "license": "BSD-3-Clause"
     },
+    "node_modules/js-md5": {
+      "version": "0.8.3",
+      "resolved": "https://registry.npmjs.org/js-md5/-/js-md5-0.8.3.tgz",
+      "integrity": "sha512-qR0HB5uP6wCuRMrWPTrkMaev7MJZwJuuw4fnwAzRgP4J4/F8RwtodOKpGp4XpqsLBFzzgqIO42efFAyz2Et6KQ==",
+      "license": "MIT"
+    },
     "node_modules/js-tokens": {
       "version": "4.0.0",
       "resolved": "https://registry.npmjs.org/js-tokens/-/js-tokens-4.0.0.tgz",
@@ -12336,6 +12482,25 @@
         "url": "https://github.com/sponsors/antonk52"
       }
     },
+    "node_modules/linebreak": {
+      "version": "1.1.0",
+      "resolved": "https://registry.npmjs.org/linebreak/-/linebreak-1.1.0.tgz",
+      "integrity": "sha512-MHp03UImeVhB7XZtjd0E4n6+3xr5Dq/9xI/5FptGk5FrbDR3zagPa2DS6U8ks/3HjbKWG9Q1M2ufOzxV2qLYSQ==",
+      "license": "MIT",
+      "dependencies": {
+        "base64-js": "0.0.8",
+        "unicode-trie": "^2.0.0"
+      }
+    },
+    "node_modules/linebreak/node_modules/base64-js": {
+      "version": "0.0.8",
+      "resolved": "https://registry.npmjs.org/base64-js/-/base64-js-0.0.8.tgz",
+      "integrity": "sha512-3XSA2cR/h/73EzlXXdU6YNycmYI7+kicTxks4eJg2g39biHR84slg2+des+p7iHYhbRg/udIS4TD53WabcOUkw==",
+      "license": "MIT",
+      "engines": {
+        "node": ">= 0.4"
+      }
+    },
     "node_modules/lines-and-columns": {
       "version": "1.2.4",
       "resolved": "https://registry.npmjs.org/lines-and-columns/-/lines-and-columns-1.2.4.tgz",
@@ -12876,7 +13041,6 @@
       "version": "1.0.1",
       "resolved": "https://registry.npmjs.org/minimalistic-assert/-/minimalistic-assert-1.0.1.tgz",
       "integrity": "sha512-UtJcAD4yEaGtjPezWuO9wC4nwUnVH/8/Im3yEHQP4b67cXlD/Qr9hdITCU1xDbSEXg2XKNaP8jsReV7vQd00/A==",
-      "dev": true,
       "license": "ISC"
     },
     "node_modules/minimatch": {
@@ -13997,6 +14161,20 @@
         "through": "~2.3"
       }
     },
+    "node_modules/pdfkit": {
+      "version": "0.19.1",
+      "resolved": "https://registry.npmjs.org/pdfkit/-/pdfkit-0.19.1.tgz",
+      "integrity": "sha512-6Gzk+wDwTs4VSxsR5rCMTnIl5nlmkye1oWB0l2hDB1EX6ZNSIBroKQEv+2+fPPn+stVjyqzmsqRJVDfB9fo5DA==",
+      "license": "MIT",
+      "dependencies": {
+        "@noble/ciphers": "^1.0.0",
+        "@noble/hashes": "^1.6.0",
+        "fontkit": "^2.0.4",
+        "js-md5": "^0.8.3",
+        "linebreak": "^1.1.0",
+        "png-js": "^1.1.0"
+      }
+    },
     "node_modules/pend": {
       "version": "1.2.0",
       "resolved": "https://registry.npmjs.org/pend/-/pend-1.2.0.tgz",
@@ -14242,6 +14420,14 @@
       "integrity": "sha512-fnWVljUchTro6RiCFvCXBbNhJc2NijN7oIQxbwsyL0buWJPG85v81ehlHI9fXrJsMNgTofEoWIQeClKpgxFLrg==",
       "license": "MIT"
     },
+    "node_modules/png-js": {
+      "version": "1.1.0",
+      "resolved": "https://registry.npmjs.org/png-js/-/png-js-1.1.0.tgz",
+      "integrity": "sha512-PM/uYGzGdNSzqeOgly68+6wKQDL1SY0a/N+OEa/+br6LnHWOAJB0Npiamnodfq3jd2LS/i2fMeOKSAILjA+m5Q==",
+      "dependencies": {
+        "browserify-zlib": "^0.2.0"
+      }
+    },
     "node_modules/possible-typed-array-names": {
       "version": "1.1.0",
       "resolved": "https://registry.npmjs.org/possible-typed-array-names/-/possible-typed-array-names-1.1.0.tgz",
@@ -15772,6 +15958,12 @@
         "node": ">=8"
       }
     },
+    "node_modules/restructure": {
+      "version": "3.0.2",
+      "resolved": "https://registry.npmjs.org/restructure/-/restructure-3.0.2.tgz",
+      "integrity": "sha512-gSfoiOEA0VPE6Tukkrr7I0RBdE0s7H1eFCDBk05l1KIQT1UIKNc5JZy6jdyW6eYH3aR3g5b3PuL77rq0hvwtAw==",
+      "license": "MIT"
+    },
     "node_modules/retry": {
       "version": "0.13.1",
       "resolved": "https://registry.npmjs.org/retry/-/retry-0.13.1.tgz",
@@ -17637,6 +17829,12 @@
       "dev": true,
       "license": "MIT"
     },
+    "node_modules/tiny-inflate": {
+      "version": "1.0.3",
+      "resolved": "https://registry.npmjs.org/tiny-inflate/-/tiny-inflate-1.0.3.tgz",
+      "integrity": "sha512-pkY1fj1cKHb2seWDy0B16HeWyczlJA9/WW3u3c4z/NiWDsO3DOU5D7nhTLE9CF0yXv/QZFY7sEJmj24dK+Rrqw==",
+      "license": "MIT"
+    },
     "node_modules/tmp": {
       "version": "0.2.7",
       "resolved": "https://registry.npmjs.org/tmp/-/tmp-0.2.7.tgz",
@@ -18353,7 +18551,6 @@
       "version": "7.19.2",
       "resolved": "https://registry.npmjs.org/undici-types/-/undici-types-7.19.2.tgz",
       "integrity": "sha512-qYVnV5OEm2AW8cJMCpdV20CDyaN3g0AjDlOGf1OW4iaDEx8MwdtChUp4zu4H0VP3nDRF/8RKWH+IPp9uW0YGZg==",
-      "dev": true,
       "license": "MIT"
     },
     "node_modules/unescape": {
@@ -18412,6 +18609,16 @@
         "node": ">=4"
       }
     },
+    "node_modules/unicode-properties": {
+      "version": "1.4.1",
+      "resolved": "https://registry.npmjs.org/unicode-properties/-/unicode-properties-1.4.1.tgz",
+      "integrity": "sha512-CLjCCLQ6UuMxWnbIylkisbRj31qxHPAurvena/0iwSVbQ2G1VY5/HjV0IRabOEbDHlzZlRdCrD4NhB0JtU40Pg==",
+      "license": "MIT",
+      "dependencies": {
+        "base64-js": "^1.3.0",
+        "unicode-trie": "^2.0.0"
+      }
+    },
     "node_modules/unicode-property-aliases-ecmascript": {
       "version": "2.2.0",
       "resolved": "https://registry.npmjs.org/unicode-property-aliases-ecmascript/-/unicode-property-aliases-ecmascript-2.2.0.tgz",
@@ -18422,6 +18629,22 @@
         "node": ">=4"
       }
     },
+    "node_modules/unicode-trie": {
+      "version": "2.0.0",
+      "resolved": "https://registry.npmjs.org/unicode-trie/-/unicode-trie-2.0.0.tgz",
+      "integrity": "sha512-x7bc76x0bm4prf1VLg79uhAzKw8DVboClSN5VxJuQ+LKDOVEW9CdH+VY7SP+vX7xCYQqzzgQpFqz15zeLvAtZQ==",
+      "license": "MIT",
+      "dependencies": {
+        "pako": "^0.2.5",
+        "tiny-inflate": "^1.0.0"
+      }
+    },
+    "node_modules/unicode-trie/node_modules/pako": {
+      "version": "0.2.9",
+      "resolved": "https://registry.npmjs.org/pako/-/pako-0.2.9.tgz",
+      "integrity": "sha512-NUcwaKxUxWrZLpDG+z/xZaCgQITkA/Dv4V/T6bw7VON6l1Xz/VnrBqrYjZQ12TamKHzITTfOEIYUj48y2KXImA==",
+      "license": "MIT"
+    },
     "node_modules/unicorn-magic": {
       "version": "0.3.0",
       "resolved": "https://registry.npmjs.org/unicorn-magic/-/unicorn-magic-0.3.0.tgz",
@@ -19363,6 +19586,24 @@
         }
       }
     },
+    "node_modules/xml": {
+      "version": "1.0.1",
+      "resolved": "https://registry.npmjs.org/xml/-/xml-1.0.1.tgz",
+      "integrity": "sha512-huCv9IH9Tcf95zuYCsQraZtWnJvBtLVE0QHMOs8bWyZAFZNDcYjsPq1nEx8jKA9y+Beo9v+7OBPRisQTjinQMw==",
+      "license": "MIT"
+    },
+    "node_modules/xml-js": {
+      "version": "1.6.11",
+      "resolved": "https://registry.npmjs.org/xml-js/-/xml-js-1.6.11.tgz",
+      "integrity": "sha512-7rVi2KMfwfWFl+GpPg6m80IVMWXLRjO+PxTq7V2CDhoGak0wzYzFgUY2m4XJ47OGdXd8eLE8EmwfAmdjw7lC1g==",
+      "license": "MIT",
+      "dependencies": {
+        "sax": "^1.2.4"
+      },
+      "bin": {
+        "xml-js": "bin/cli.js"
+      }
+    },
     "node_modules/xml-name-validator": {
       "version": "5.0.0",
       "resolved": "https://registry.npmjs.org/xml-name-validator/-/xml-name-validator-5.0.0.tgz",
diff --git a/package.json b/package.json
index 8a84222e..d7aed3e4 100644
--- a/package.json
+++ b/package.json
@@ -39,6 +39,7 @@
     "test:worker:crm": "npm run db:smoke-seed && npm run build:worker && node scripts/crm-worker-integration-test.js",
     "test:worker:commerce": "npm run db:smoke-seed && npm run build:worker && node scripts/commerce-worker-integration-test.js",
     "test:worker:assets": "npm run db:smoke-seed && npm run build:worker && node scripts/asset-worker-integration-test.js",
+    "test:worker:exports": "npm run db:smoke-seed && npm run build:worker && node scripts/export-worker-integration-test.js",
     "test:worker:imports": "npm run db:smoke-seed && npm run build:worker && node scripts/import-worker-integration-test.js",
     "test:worker:public-banks": "npm run db:smoke-seed && npm run build:worker && node scripts/public-bank-worker-integration-test.js",
     "test:readiness": "node scripts/production-readiness-check-test.js",
diff --git a/scripts/api-integration-test.js b/scripts/api-integration-test.js
index e39175be..c0dbf517 100644
--- a/scripts/api-integration-test.js
+++ b/scripts/api-integration-test.js
@@ -2814,6 +2814,45 @@ async function testTenantContentAdmin() {
   const exportedQuestionWithAnswer = paperExport.export?.questions?.find(item => item.id === question.item.id);
   assert.equal(exportedQuestionWithAnswer?.answerText, '4', 'paper export should include answers when requested');
 
+  const pdfExportJob = await request('/api/tenant-content/exports/questions', {
+    userId: TENANT_ADMIN_USER_ID,
+    method: 'POST',
+    body: {
+      scopeType: 'collection',
+      scopeId: collection.item.id,
+      format: 'pdf',
+      exportType: 'paper',
+      includeAnswers: false,
+      includeExplanations: false,
+      options: {
+        title: '集成测试 PDF 异步导出',
+        watermarkText: '仅供内部测试',
+        publishToAssets: true,
+        assetVisibility: 'tenant',
+      },
+    },
+  });
+  assert.equal(pdfExportJob.job?.status, 'pending', 'PDF export should create a pending async export job');
+  assert.equal(pdfExportJob.export, null, 'PDF export should not return inline binary payload');
+
+  const docxExportJob = await request('/api/tenant-content/exports/questions', {
+    userId: TENANT_ADMIN_USER_ID,
+    method: 'POST',
+    body: {
+      scopeType: 'collection',
+      scopeId: collection.item.id,
+      format: 'docx',
+      exportType: 'paper',
+      includeAnswers: true,
+      includeExplanations: true,
+      options: {
+        title: '集成测试 Word 异步导出',
+        watermarkText: '仅供内部测试',
+      },
+    },
+  });
+  assert.equal(docxExportJob.job?.status, 'pending', 'DOCX export should create a pending async export job');
+
   const partnerExportDenied = await request('/api/tenant-content/exports/questions', {
     tenantId: PARTNER_TENANT_ID,
     userId: PARTNER_TENANT_ADMIN_USER_ID,
@@ -2831,6 +2870,10 @@ async function testTenantContentAdmin() {
     exportJobs.items?.some(item => item.id === exportNoAnswers.job.id && item.questionCount === 1),
     'question export jobs should list completed export audit records',
   );
+  const listedPdfJob = exportJobs.items?.find(item => item.id === pdfExportJob.job.id);
+  assert.equal(listedPdfJob?.status, 'pending', 'question export jobs should expose pending binary export jobs');
+  assert.equal(listedPdfJob?.assetId, null, 'pending binary export job should not have an asset yet');
+  assert.equal(listedPdfJob?.attemptCount, 0, 'pending binary export job should expose worker attempt count');
   const invalidExportJobsScope = await request('/api/tenant-content/exports/jobs', {
     userId: TENANT_ADMIN_USER_ID,
     query: { scopeType: 'not_supported' },
diff --git a/scripts/export-worker-integration-test.js b/scripts/export-worker-integration-test.js
new file mode 100644
index 00000000..0a668fdb
--- /dev/null
+++ b/scripts/export-worker-integration-test.js
@@ -0,0 +1,249 @@
+import assert from 'node:assert/strict';
+import fs from 'node:fs/promises';
+import path from 'node:path';
+import pg from 'pg';
+import { spawn } from 'node:child_process';
+import JSZip from 'jszip';
+
+const databaseUrl = process.env.DATABASE_URL || 'postgresql://postgres:postgres@127.0.0.1:54322/postgres';
+const tenantId = '00000000-0000-0000-0000-000000000001';
+const tenantAdminUserId = '00000000-0000-0000-0000-000000000102';
+const collectionId = '00000000-0000-0000-0000-000000000615';
+const localStorageRoot = '.local-storage-test';
+
+const ids = {
+  pdfJob: '21000000-0000-0000-0000-000000000901',
+  docxJob: '21000000-0000-0000-0000-000000000902',
+  noAnswerDocxJob: '21000000-0000-0000-0000-000000000903',
+};
+
+const jobIds = [ids.pdfJob, ids.docxJob, ids.noAnswerDocxJob];
+
+async function runWorkerOnce() {
+  const child = spawn(process.execPath, ['apps/worker/dist/apps/worker/src/index.js', '--once', '--job', 'exports'], {
+    cwd: process.cwd(),
+    env: {
+      ...process.env,
+      DATABASE_URL: databaseUrl,
+      WORKER_EXPORT_BATCH_SIZE: '20',
+      WORKER_EXPORT_ID: 'export-worker-test',
+      STORAGE_DEFAULT_PROVIDER: 'local_dev',
+      STORAGE_DEFAULT_BUCKET: 'tenant-assets',
+      EXPORT_LOCAL_STORAGE_ROOT: localStorageRoot,
+      STORAGE_REQUIRE_TENANT_PREFIX: 'true',
+    },
+    stdio: ['ignore', 'pipe', 'pipe'],
+    windowsHide: true,
+  });
+  let output = '';
+  child.stdout.on('data', chunk => {
+    output += chunk.toString();
+  });
+  child.stderr.on('data', chunk => {
+    output += chunk.toString();
+  });
+  const code = await new Promise(resolve => child.on('exit', resolve));
+  assert.equal(code, 0, `worker should exit 0\n${output}`);
+  assert.match(output, /exports batch processed=\d+/, 'worker output should include exports summary');
+  return output;
+}
+
+async function cleanup(pool) {
+  await pool.query(
+    `
+    delete from public.audit_logs
+    where tenant_id = $1
+      and (
+        (target_type = 'content_export_job' and target_id = any($2::text[]))
+        or (target_type = 'content_asset' and details->>'exportJobId' = any($2::text[]))
+      )
+    `,
+    [tenantId, jobIds],
+  );
+  await pool.query(
+    `
+    delete from public.content_assets
+    where tenant_id = $1
+      and metadata->>'exportJobId' = any($2::text[])
+    `,
+    [tenantId, jobIds],
+  );
+  await pool.query(
+    `
+    delete from public.content_export_jobs
+    where tenant_id = $1 and id = any($2::uuid[])
+    `,
+    [tenantId, jobIds],
+  );
+  await fs.rm(path.resolve(process.cwd(), localStorageRoot), { recursive: true, force: true });
+}
+
+async function seed(pool) {
+  await pool.query(
+    `
+    insert into public.content_export_jobs (
+      id, tenant_id, created_by, export_type, format, scope_type, scope_id,
+      status, include_answers, include_explanations, include_video_refs,
+      question_count, summary, options, output_metadata, created_at, updated_at
+    )
+    values
+      (
+        $1, $4, $5, 'paper', 'pdf', 'collection', $6,
+        'pending', false, false, false,
+        0, '{"source":"export-worker-test"}'::jsonb,
+        '{"title":"导出测试无答案试卷","watermarkText":"内部水印","publishToAssets":true,"assetVisibility":"tenant","limit":50}'::jsonb,
+        '{"delivery":"content_asset_pending","maxQuestionLimit":50}'::jsonb,
+        now() - interval '1 minute', now() - interval '1 minute'
+      ),
+      (
+        $2, $4, $5, 'paper', 'docx', 'collection', $6,
+        'pending', true, true, false,
+        0, '{"source":"export-worker-test"}'::jsonb,
+        '{"title":"导出测试含答案文档","watermarkText":"内部水印","publishToAssets":false,"assetVisibility":"private","limit":50}'::jsonb,
+        '{"delivery":"content_asset_pending","maxQuestionLimit":50}'::jsonb,
+        now() - interval '1 minute', now() - interval '1 minute'
+      ),
+      (
+        $3, $4, $5, 'paper', 'docx', 'collection', $6,
+        'pending', false, false, false,
+        0, '{"source":"export-worker-test"}'::jsonb,
+        '{"title":"导出测试无答案文档","watermarkText":"内部水印","publishToAssets":false,"assetVisibility":"private","limit":50}'::jsonb,
+        '{"delivery":"content_asset_pending","maxQuestionLimit":50}'::jsonb,
+        now() - interval '1 minute', now() - interval '1 minute'
+      )
+    `,
+    [ids.pdfJob, ids.docxJob, ids.noAnswerDocxJob, tenantId, tenantAdminUserId, collectionId],
+  );
+}
+
+function countFromWorkerOutput(output, name) {
+  const match = output.match(new RegExp(`${name}=(\\d+)`));
+  return match ? Number(match[1]) : 0;
+}
+
+async function generatedFileBody(asset) {
+  const fullPath = path.resolve(process.cwd(), localStorageRoot, asset.bucket, asset.object_key);
+  return fs.readFile(fullPath);
+}
+
+async function assertGeneratedFile(asset) {
+  const body = await generatedFileBody(asset);
+  assert.equal(body.length, Number(asset.file_size_bytes), 'generated file size should match content_assets');
+  assert.ok(body.length > 500, 'generated binary export should not be empty');
+  if (asset.mime_type === 'application/pdf') {
+    assert.equal(body.subarray(0, 4).toString('utf8'), '%PDF', 'PDF export should have PDF header');
+  } else {
+    assert.equal(body.subarray(0, 2).toString('hex'), '504b', 'DOCX export should be a zipped Office file');
+  }
+}
+
+async function docxText(asset) {
+  const zip = await JSZip.loadAsync(await generatedFileBody(asset));
+  const xml = await zip.file('word/document.xml')?.async('string');
+  assert.ok(xml, 'DOCX should contain word/document.xml');
+  return xml
+    .replace(/<[^>]+>/g, '')
+    .replace(/&lt;/g, '<')
+    .replace(/&gt;/g, '>')
+    .replace(/&amp;/g, '&');
+}
+
+async function main() {
+  const pool = new pg.Pool({ connectionString: databaseUrl });
+  let seeded = false;
+  try {
+    await pool.query('begin');
+    await cleanup(pool);
+    await seed(pool);
+    await pool.query('commit');
+    seeded = true;
+
+    const output = await runWorkerOnce();
+    assert.ok(countFromWorkerOutput(output, 'completed') >= 3, 'worker should complete all binary export jobs');
+
+    const jobs = await pool.query(
+      `
+      select id, status, asset_id, question_count, output_metadata
+      from public.content_export_jobs
+      where tenant_id = $1 and id = any($2::uuid[])
+      order by id
+      `,
+      [tenantId, jobIds],
+    );
+    assert.equal(jobs.rowCount, 3, 'all export jobs should still exist');
+    for (const job of jobs.rows) {
+      assert.equal(job.status, 'completed', 'export job should be completed');
+      assert.ok(job.asset_id, 'completed export job should reference a content asset');
+      assert.equal(job.output_metadata?.delivery, 'content_asset', 'export job should switch delivery to content_asset');
+      assert.ok(job.output_metadata?.checksumSha256, 'export job should record output checksum');
+      assert.ok(Number(job.question_count) >= 1, 'export job should record question count');
+    }
+
+    const assets = await pool.query(
+      `
+      select id, asset_type, title, file_name, storage_provider, bucket,
+             object_key, mime_type, file_size_bytes, checksum_sha256,
+             upload_status, visibility, status, metadata
+      from public.content_assets
+      where tenant_id = $1 and metadata->>'exportJobId' = any($2::text[])
+      order by file_name
+      `,
+      [tenantId, jobIds],
+    );
+    assert.equal(assets.rowCount, 3, 'worker should create three export assets');
+    const pdfAsset = assets.rows.find(asset => asset.metadata?.exportJobId === ids.pdfJob);
+    const docxAsset = assets.rows.find(asset => asset.metadata?.exportJobId === ids.docxJob);
+    const noAnswerDocxAsset = assets.rows.find(asset => asset.metadata?.exportJobId === ids.noAnswerDocxJob);
+    assert.ok(pdfAsset, 'PDF export should create a PDF content asset');
+    assert.ok(docxAsset, 'DOCX export should create a Word content asset');
+    assert.ok(noAnswerDocxAsset, 'no-answer DOCX export should create a Word content asset');
+    assert.equal(pdfAsset.mime_type, 'application/pdf', 'PDF asset should use application/pdf');
+    assert.ok(docxAsset.mime_type?.includes('wordprocessingml'), 'DOCX asset should use Word MIME type');
+    assert.equal(pdfAsset.upload_status, 'verified', 'PDF asset should be verified');
+    assert.equal(docxAsset.upload_status, 'verified', 'DOCX asset should be verified');
+    assert.equal(noAnswerDocxAsset.upload_status, 'verified', 'no-answer DOCX asset should be verified');
+    assert.equal(pdfAsset.visibility, 'tenant', 'published PDF export should use requested asset visibility');
+    assert.equal(docxAsset.visibility, 'private', 'non-published DOCX export should remain private');
+    assert.equal(noAnswerDocxAsset.visibility, 'private', 'no-answer DOCX export should remain private');
+    await assertGeneratedFile(pdfAsset);
+    await assertGeneratedFile(docxAsset);
+    await assertGeneratedFile(noAnswerDocxAsset);
+
+    const answeredDocxText = await docxText(docxAsset);
+    assert.match(answeredDocxText, /答案/, 'DOCX with answers should include answer labels');
+    assert.match(answeredDocxText, /后端需要统一校验租户/, 'DOCX with answers should include subjective answer text');
+    const noAnswerDocxText = await docxText(noAnswerDocxAsset);
+    assert.doesNotMatch(noAnswerDocxText, /后端需要统一校验租户/, 'DOCX without answers should not leak subjective answer text');
+    assert.doesNotMatch(noAnswerDocxText, /基础加法。/, 'DOCX without explanations should not leak explanations');
+
+    const audits = await pool.query(
+      `
+      select action, details
+      from public.audit_logs
+      where tenant_id = $1
+        and target_type = 'content_export_job'
+        and target_id = any($2::text[])
+      `,
+      [tenantId, jobIds],
+    );
+    assert.ok(
+      audits.rows.some(row => row.action === 'content.questions.export_rendered' && row.details?.assetId),
+      'worker should write rendered audit log',
+    );
+
+    console.log('Export worker integration test complete.');
+  } catch (error) {
+    await pool.query('rollback').catch(() => {});
+    throw error;
+  } finally {
+    if (seeded) {
+      await cleanup(pool).catch(() => {});
+    }
+    await pool.end();
+  }
+}
+
+main().catch(error => {
+  console.error(error);
+  process.exit(1);
+});
diff --git a/supabase/migrations/202606290018_content_export_binary_worker.sql b/supabase/migrations/202606290018_content_export_binary_worker.sql
new file mode 100644
index 00000000..cd0f9447
--- /dev/null
+++ b/supabase/migrations/202606290018_content_export_binary_worker.sql
@@ -0,0 +1,40 @@
+alter table public.content_export_jobs
+  drop constraint if exists content_export_jobs_format_check,
+  drop constraint if exists content_export_jobs_status_check;
+
+alter table public.content_export_jobs
+  add constraint content_export_jobs_format_check
+  check (format in ('json', 'paper_json', 'print_payload', 'pdf', 'docx'));
+
+alter table public.content_export_jobs
+  add constraint content_export_jobs_status_check
+  check (status in ('pending', 'rendering', 'completed', 'failed'));
+
+alter table public.content_export_jobs
+  alter column status set default 'pending';
+
+alter table public.content_export_jobs
+  add column if not exists asset_id uuid references public.content_assets(id) on delete set null,
+  add column if not exists locked_at timestamptz,
+  add column if not exists locked_by text,
+  add column if not exists attempt_count integer not null default 0,
+  add column if not exists max_attempts integer not null default 3,
+  add column if not exists next_attempt_at timestamptz,
+  add column if not exists started_at timestamptz;
+
+do $$
+begin
+  if not exists (select 1 from pg_constraint where conname = 'content_export_jobs_attempt_count_check') then
+    alter table public.content_export_jobs
+      add constraint content_export_jobs_attempt_count_check
+      check (attempt_count >= 0 and max_attempts > 0 and attempt_count <= max_attempts);
+  end if;
+end $$;
+
+create index if not exists idx_content_export_jobs_worker_queue
+  on public.content_export_jobs(status, next_attempt_at, created_at)
+  where status = 'pending';
+
+create index if not exists idx_content_export_jobs_asset
+  on public.content_export_jobs(tenant_id, asset_id)
+  where asset_id is not null;