Add self-service GDPR data export (#68)

A signed-in account can export all of its own data — profile, a list of
its memberships/grants, and the Markdown of its personal pond plus the
shared ponds it owns — as one ZIP. Foreign content never appears: only
owned ponds are bundled and the per-page read filter (reused from #65)
runs for each.

- Reuse the conversion-job queue as the async carrier: a `data_export`
  job whose worker branch resolves DataExportService via a token (no DI
  cycle), builds the ZIP, and stores it with an `expiresAt`. The download
  link 404s past expiry and an hourly scheduled purge drops the bytes
  (data minimization, security.md §Privacy).
- Extract ExportService.appendPondMarkdown so the pond ZIP (#65) and the
  data export share one read-filtered pond archiver.
- Rate-limit requests per account (RateLimitService); POST
  /users/me/data-export enqueues, GET /jobs/:id(/result) poll/download.
- Settings UI "Export my data" (de+en); web share pollJob/downloadJobResult
  between the document and data export hooks.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EwZ4jR4KFAPvpjWevfUGX1

apps/api/prisma/migrations/20260710120000_data_export_expiry/migration.sql

@@ -0,0 +1,2 @@
+-- AlterTable
+ALTER TABLE "conversion_jobs" ADD COLUMN     "expires_at" TIMESTAMP(3);

apps/api/prisma/schema.prisma

@@ -550,6 +550,10 @@ model ConversionJob {
   resultMimeType String?             @map("result_mime_type")
   errorCode      String?             @map("error_code")
   lockedAt       DateTime?           @map("locked_at")
+  /// For a data-export job (#68): when its stored result stops being
+  /// downloadable and is purged (GDPR data minimization). Null for every
+  /// other job kind, whose result never expires.
+  expiresAt      DateTime?           @map("expires_at")
   createdAt      DateTime            @default(now()) @map("created_at")
   updatedAt      DateTime            @updatedAt @map("updated_at")
 

apps/api/src/import-export/conversion-job.service.ts

@@ -89,6 +89,9 @@ export class ConversionJobService {
   async resultForOwner(id: string, userId: string): Promise<ConversionResultPayload> {
     const job = await this.ownedJob(id, userId);
     if (job.status !== 'SUCCEEDED' || !job.result) throw new NotFoundException();
+    // A data-export link expires (#68): past its window the result is treated as
+    // gone (a scheduled purge deletes the bytes), so the download 404s.
+    if (job.expiresAt && job.expiresAt.getTime() < Date.now()) throw new NotFoundException();
     return {
       bytes: Buffer.from(job.result),
       mimeType: job.resultMimeType ?? 'application/octet-stream',
@@ -111,6 +114,8 @@ export class ConversionJobService {
       errorCode: job.errorCode,
       // Set once an import job succeeds (#63) so the client can open the page.
       resultPageId: job.resultPageId,
+      // Set for a data-export job (#68): when its download link stops working.
+      expiresAt: job.expiresAt?.toISOString() ?? null,
       createdAt: job.createdAt.toISOString(),
       updatedAt: job.updatedAt.toISOString(),
     };

apps/api/src/import-export/conversion-worker.service.ts

@@ -6,6 +6,12 @@ import { PinoLogger } from 'nestjs-pino';
 import { AppConfig } from '../config/app-config.service';
 import { PrismaService } from '../prisma/prisma.service';
 
+import {
+  DATA_EXPORT_KIND,
+  DATA_EXPORT_PROCESSOR,
+  DATA_EXPORT_TTL_MS,
+  DataExportProcessor,
+} from './data-export.constants';
 import { GotenbergRenderer, RenderError } from './gotenberg.renderer';
 import { IMPORT_PROCESSOR, ImportProcessor, isImportKind } from './import.constants';
 import { ConversionError, ConversionResult, PandocConverter } from './pandoc.converter';
@@ -127,15 +133,26 @@ export class ConversionWorker implements OnModuleInit, OnModuleDestroy {
         await this.moduleRef.get<ImportProcessor>(IMPORT_PROCESSOR, { strict: false }).run(job);
         return;
       }
-      // PDF export renders HTML through Gotenberg (#67); everything else is a
-      // pandoc byte→byte conversion (#62/#65).
-      const output =
-        job.targetFormat === 'pdf'
-          ? {
-              bytes: await this.renderer.renderHtmlToPdf(Buffer.from(job.input).toString('utf8')),
-              mimeType: 'application/pdf',
-            }
-          : await this.convert(job);
+      // A data export gathers the account's own data into a ZIP (#68) whose
+      // download link expires; a PDF export renders HTML through Gotenberg
+      // (#67); everything else is a pandoc byte→byte conversion (#62/#65).
+      let output: { bytes: Buffer; mimeType: string };
+      let expiresAt: Date | null = null;
+      if (job.kind === DATA_EXPORT_KIND) {
+        // Resolved lazily via a token so this file never imports the export
+        // service's (avoids a construction cycle), mirroring the import path.
+        output = await this.moduleRef
+          .get<DataExportProcessor>(DATA_EXPORT_PROCESSOR, { strict: false })
+          .build(job);
+        expiresAt = new Date(Date.now() + DATA_EXPORT_TTL_MS);
+      } else if (job.targetFormat === 'pdf') {
+        output = {
+          bytes: await this.renderer.renderHtmlToPdf(Buffer.from(job.input).toString('utf8')),
+          mimeType: 'application/pdf',
+        };
+      } else {
+        output = await this.convert(job);
+      }
       await this.prisma.conversionJob.update({
         where: { id: job.id },
         data: {
@@ -144,6 +161,7 @@ export class ConversionWorker implements OnModuleInit, OnModuleDestroy {
           result: new Uint8Array(output.bytes),
           resultMimeType: output.mimeType,
           errorCode: null,
+          expiresAt,
         },
       });
       this.logger.info(

apps/api/src/import-export/data-export.constants.ts

@@ -0,0 +1,23 @@
+import type { ConversionJob } from '@prisma/client';
+
+/**
+ * Wiring for the GDPR data export (issue #68). The conversion worker builds a
+ * `data_export` job through this token — never by importing the service — so
+ * the worker's file stays free of a construction cycle (mirrors the import
+ * pipeline's {@link ./import.constants.ts}).
+ */
+export const DATA_EXPORT_PROCESSOR = Symbol('DATA_EXPORT_PROCESSOR');
+
+/** The {@link ConversionJob.kind} of a self-service account data export. */
+export const DATA_EXPORT_KIND = 'data_export';
+
+/** How long a generated export stays downloadable before its bytes are purged
+ * (security.md §Privacy: data minimization — the archive is a transient copy,
+ * not stored indefinitely). */
+export const DATA_EXPORT_TTL_MS = 24 * 60 * 60 * 1000;
+
+/** Builds the ZIP for a `data_export` job. The worker resolves this behind the
+ * token above and stores the returned bytes as the job result. */
+export interface DataExportProcessor {
+  build(job: ConversionJob): Promise<{ bytes: Buffer; mimeType: string }>;
+}

apps/api/src/import-export/data-export.controller.ts

@@ -0,0 +1,24 @@
+import { Controller, Post, Req } from '@nestjs/common';
+import { ConversionJobView } from '@dorfteich/shared';
+
+import { AuthedRequest } from '../auth/auth.guard';
+import { AuthenticatedOnly } from '../permissions/permission.decorators';
+
+import { DataExportService } from './data-export.service';
+
+/**
+ * Self-service GDPR data export (issue #68, security.md §Privacy). A signed-in
+ * account requests its own export here; the job is owner-scoped, so polling and
+ * download stay on the generic `GET /jobs/:id(/result)` routes (JobsController).
+ */
+@AuthenticatedOnly()
+@Controller('users/me')
+export class DataExportController {
+  constructor(private readonly dataExport: DataExportService) {}
+
+  /** Enqueue an export of the caller's own data; rate-limited per account. */
+  @Post('data-export')
+  request(@Req() request: AuthedRequest): Promise<ConversionJobView> {
+    return this.dataExport.request(request.user!);
+  }
+}

apps/api/src/import-export/data-export.service.db.test.ts

@@ -0,0 +1,299 @@
+import { INestApplication } from '@nestjs/common';
+import { PrismaClient } from '@prisma/client';
+import { unzipSync } from 'fflate';
+import request from 'supertest';
+import { afterAll, beforeAll, beforeEach, describe, expect, it } from 'vitest';
+
+import { AuthTokensService } from '../auth/auth-tokens.service';
+import { FilesService } from '../files/files.service';
+import { createTestApp, sessionCookieOf } from '../testing/test-app';
+import { createTestPrisma, grantOwnerAdmin, hasTestDb, uniqueSuffix } from '../testing/test-db';
+import { UsersService } from '../users/users.service';
+
+import { ConversionWorker } from './conversion-worker.service';
+import { DataExportService } from './data-export.service';
+
+/** GDPR data export (issue #68): a signed-in account assembles its own profile,
+ * memberships, and the Markdown of every pond it owns into one ZIP — with a
+ * download link that expires — and never another user's content. */
+
+const PNG_BASE64 =
+  'iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAC0lEQVR42mP8z8BQDwAEhQGAhKmMIQAAAABJRU5ErkJggg==';
+
+function zipEntries(buffer: Buffer): Record<string, Uint8Array> {
+  return unzipSync(new Uint8Array(buffer));
+}
+
+function textOf(entries: Record<string, Uint8Array>, name: string): string {
+  return Buffer.from(entries[name]!).toString('utf8');
+}
+
+async function downloadZip(server: unknown, path: string, cookie: string): Promise<Buffer> {
+  const res = await request(server as never)
+    .get(path)
+    .set('Cookie', cookie)
+    .buffer(true)
+    .parse((r, cb) => {
+      const chunks: Buffer[] = [];
+      r.on('data', (c: Buffer) => chunks.push(c));
+      r.on('end', () => cb(null, Buffer.concat(chunks)));
+    })
+    .expect(200);
+  return res.body as Buffer;
+}
+
+describe.skipIf(!hasTestDb)('data export (e2e, issue #68)', () => {
+  let app: INestApplication;
+  let prisma: PrismaClient;
+  let worker: ConversionWorker;
+  let files: FilesService;
+  const suffix = uniqueSuffix();
+  const password = 'exportiere alle meine daten 1';
+
+  let ownerId: string;
+  let ownerCookie: string;
+  let personalSlug: string;
+  let ownedSharedSlug: string;
+  let foreignSlug: string;
+
+  const api = () => request(app.getHttpServer());
+
+  async function login(username: string): Promise<string> {
+    const res = await api()
+      .post('/api/v1/auth/login')
+      .send({ usernameOrEmail: username, password })
+      .expect(200);
+    return sessionCookieOf(res);
+  }
+
+  async function createUser(username: string): Promise<string> {
+    const user = await app.get(UsersService).createUser({
+      username,
+      email: `${username}@example.org`,
+      displayName: username,
+      password,
+      locale: 'en',
+    });
+    // Verify via the endpoint (not a direct flag flip) so the personal pond and
+    // its owner-admin grant are created exactly as in production (#52).
+    const token = await app.get(AuthTokensService).issue(user.id, 'EMAIL_VERIFICATION', 600);
+    await api().post('/api/v1/auth/verify-email').send({ token }).expect(204);
+    return user.id;
+  }
+
+  async function seedPage(pondId: string, title: string, markdown: string): Promise<void> {
+    await prisma.page.create({
+      data: {
+        pondId,
+        title,
+        slug: title.toLowerCase().replace(/[^a-z0-9]+/g, '-'),
+        ydocState: new Uint8Array(),
+        sortKey: title,
+        createdBy: ownerId,
+        contentCache: { create: { plainText: markdown, markdown, html: '', outline: [] } },
+      },
+    });
+  }
+
+  beforeAll(async () => {
+    prisma = createTestPrisma();
+    await prisma.rateLimit.deleteMany({});
+    app = await createTestApp();
+    worker = app.get(ConversionWorker);
+    files = app.get(FilesService);
+
+    ownerId = await createUser(`odette-owner-${suffix}`);
+    ownerCookie = await login(`odette-owner-${suffix}`);
+
+    const personal = await prisma.pond.findFirstOrThrow({
+      where: { ownerId, type: 'PERSONAL' },
+    });
+    personalSlug = personal.slug;
+
+    // A shared pond the owner created — its content belongs in the export.
+    const shared = await prisma.pond.create({
+      data: {
+        slug: `owned-shared-${suffix}`,
+        name: 'Owned Shared',
+        type: 'SHARED',
+        ownerId,
+        usage: { create: {} },
+      },
+    });
+    ownedSharedSlug = shared.slug;
+    await grantOwnerAdmin(prisma, shared.id, ownerId);
+
+    // A stored image referenced from a personal-pond page (media inclusion).
+    const image = await files.upload({ id: ownerId } as never, personal.id, {
+      buffer: Buffer.from(PNG_BASE64, 'base64'),
+      size: 70,
+      originalname: 'dot.png',
+    });
+    await seedPage(personal.id, 'Diary', `# Diary\n\nMine. ![dot](${image.id})`);
+    await seedPage(shared.id, 'Shared Note', '# Shared Note\n\nAlso mine.');
+
+    // A foreign user's pond the owner is only a *reader* of: it must show up in
+    // memberships, but its content must never appear in the export.
+    const foreignId = await createUser(`fred-foreign-${suffix}`);
+    const foreign = await prisma.pond.create({
+      data: {
+        slug: `foreign-secret-${suffix}`,
+        name: 'Foreign Secret',
+        type: 'SHARED',
+        ownerId: foreignId,
+        usage: { create: {} },
+      },
+    });
+    foreignSlug = foreign.slug;
+    await prisma.page.create({
+      data: {
+        pondId: foreign.id,
+        title: 'Foreign Page',
+        slug: 'foreign-page',
+        ydocState: new Uint8Array(),
+        sortKey: 'Foreign Page',
+        createdBy: foreignId,
+        contentCache: {
+          create: { plainText: 'TOP SECRET', markdown: 'TOP SECRET', html: '', outline: [] },
+        },
+      },
+    });
+    await prisma.roleGrant.create({
+      data: {
+        pondId: foreign.id,
+        subjectType: 'USER',
+        subjectId: ownerId,
+        role: 'READER',
+        scopeType: 'POND',
+        effect: 'ALLOW',
+        createdBy: foreignId,
+      },
+    });
+  });
+
+  afterAll(async () => {
+    await prisma.conversionJob.deleteMany({ where: { owner: { username: { contains: suffix } } } });
+    const where = { pond: { owner: { username: { contains: suffix } } } };
+    await prisma.attachment.deleteMany({ where });
+    await prisma.roleGrant.deleteMany({
+      where: { pond: { owner: { username: { contains: suffix } } } },
+    });
+    await prisma.page.deleteMany({ where });
+    await prisma.pond.deleteMany({ where: { owner: { username: { contains: suffix } } } });
+    await prisma.rateLimit.deleteMany({});
+    await prisma.user.deleteMany({ where: { username: { contains: suffix } } });
+    await prisma.$disconnect();
+    await app.close();
+  });
+
+  // Each test starts from a clean rate-limit table so the per-account export
+  // limit (exercised deliberately in one test) never trips the others.
+  beforeEach(async () => {
+    await prisma.rateLimit.deleteMany({});
+  });
+
+  it('exports the caller profile, memberships, and owned pond content only', async () => {
+    const enqueued = await api()
+      .post('/api/v1/users/me/data-export')
+      .set('Cookie', ownerCookie)
+      .expect(201);
+    expect(enqueued.body.kind).toBe('data_export');
+    expect(enqueued.body.status).toBe('pending');
+
+    await worker.drain();
+
+    const done = await api()
+      .get(`/api/v1/jobs/${enqueued.body.id}`)
+      .set('Cookie', ownerCookie)
+      .expect(200);
+    expect(done.body.status).toBe('succeeded');
+    // The download link carries an expiry (a future timestamp).
+    expect(done.body.expiresAt).toBeTruthy();
+    expect(new Date(done.body.expiresAt).getTime()).toBeGreaterThan(Date.now());
+
+    const zip = await downloadZip(
+      app.getHttpServer(),
+      `/api/v1/jobs/${enqueued.body.id}/result`,
+      ownerCookie,
+    );
+    const entries = zipEntries(zip);
+    const names = Object.keys(entries);
+
+    // Profile: the caller's own account fields.
+    const profile = JSON.parse(textOf(entries, 'profile.json'));
+    expect(profile.username).toBe(`odette-owner-${suffix}`);
+    expect(profile.email).toBe(`odette-owner-${suffix}@example.org`);
+
+    // Personal + owned-shared pond content is present, media included.
+    expect(names).toContain(`ponds/${personalSlug}/diary.md`);
+    expect(names.some((n) => n.startsWith(`ponds/${personalSlug}/media/`))).toBe(true);
+    expect(names).toContain(`ponds/${ownedSharedSlug}/shared-note.md`);
+
+    // Memberships list the foreign pond the owner reads…
+    const memberships = JSON.parse(textOf(entries, 'memberships.json')) as {
+      pondSlug: string;
+    }[];
+    expect(memberships.map((m) => m.pondSlug)).toContain(foreignSlug);
+
+    // …but none of the foreign pond's content appears anywhere in the archive.
+    expect(names.some((n) => n.startsWith(`ponds/${foreignSlug}/`))).toBe(false);
+    for (const name of names) {
+      if (name.endsWith('.md')) expect(textOf(entries, name)).not.toContain('TOP SECRET');
+    }
+  });
+
+  it('404s the download once the link has expired', async () => {
+    const enqueued = await api()
+      .post('/api/v1/users/me/data-export')
+      .set('Cookie', ownerCookie)
+      .expect(201);
+    await worker.drain();
+    // Backdate the expiry to simulate an elapsed link.
+    await prisma.conversionJob.update({
+      where: { id: enqueued.body.id },
+      data: { expiresAt: new Date(Date.now() - 1000) },
+    });
+
+    await api()
+      .get(`/api/v1/jobs/${enqueued.body.id}/result`)
+      .set('Cookie', ownerCookie)
+      .expect(404);
+  });
+
+  it('purges the stored bytes of an expired export', async () => {
+    const enqueued = await api()
+      .post('/api/v1/users/me/data-export')
+      .set('Cookie', ownerCookie)
+      .expect(201);
+    await worker.drain();
+    await prisma.conversionJob.update({
+      where: { id: enqueued.body.id },
+      data: { expiresAt: new Date(Date.now() - 1000) },
+    });
+
+    const purged = await app.get(DataExportService).purgeExpired();
+    expect(purged).toBeGreaterThanOrEqual(1);
+    const row = await prisma.conversionJob.findUniqueOrThrow({ where: { id: enqueued.body.id } });
+    expect(row.result).toBeNull();
+  });
+
+  it('rate-limits repeated requests from the same account', async () => {
+    await createUser(`irene-idle-${suffix}`);
+    const cookie = await login(`irene-idle-${suffix}`);
+    // The limit is a few requests per window; the run past it is rejected.
+    for (let i = 0; i < 3; i += 1) {
+      await api().post('/api/v1/users/me/data-export').set('Cookie', cookie).expect(201);
+    }
+    await api().post('/api/v1/users/me/data-export').set('Cookie', cookie).expect(429);
+  });
+
+  it('keeps a data-export job owner-scoped (a foreign user cannot poll it)', async () => {
+    const enqueued = await api()
+      .post('/api/v1/users/me/data-export')
+      .set('Cookie', ownerCookie)
+      .expect(201);
+    await createUser(`sam-stranger-${suffix}`);
+    const strangerCookie = await login(`sam-stranger-${suffix}`);
+    await api().get(`/api/v1/jobs/${enqueued.body.id}`).set('Cookie', strangerCookie).expect(404);
+  });
+});

apps/api/src/import-export/data-export.service.ts

@@ -0,0 +1,174 @@
+import { HttpException, HttpStatus, Injectable } from '@nestjs/common';
+import { ConversionJobView } from '@dorfteich/shared';
+import { ConversionJob, User } from '@prisma/client';
+import archiver from 'archiver';
+import { PinoLogger } from 'nestjs-pino';
+
+import { RateLimitService } from '../rate-limit/rate-limit.service';
+import { PrismaService } from '../prisma/prisma.service';
+
+import { ConversionJobService } from './conversion-job.service';
+import { DATA_EXPORT_KIND, DataExportProcessor } from './data-export.constants';
+import { ExportService } from './export.service';
+
+/** A data export may be requested at most this many times per window per
+ * account (security.md §Privacy: access requests are self-service but must not
+ * be a cheap way to hammer the server). */
+const MAX_REQUESTS_PER_WINDOW = 3;
+const WINDOW_SECONDS = 60 * 60;
+
+/**
+ * GDPR data export (issue #68): a signed-in user assembles all of their own
+ * data — profile, a list of their memberships/grants, and the Markdown export
+ * of their personal pond and the shared ponds they own — into a single ZIP.
+ * It is produced as a conversion job (so a large account exports out of band)
+ * with a download link that expires; foreign content never appears, because
+ * only the requester's own ponds are included and the per-page read filter
+ * (reused from #65) runs for every one.
+ */
+@Injectable()
+export class DataExportService implements DataExportProcessor {
+  constructor(
+    private readonly prisma: PrismaService,
+    private readonly exports: ExportService,
+    private readonly jobs: ConversionJobService,
+    private readonly rateLimits: RateLimitService,
+    private readonly logger: PinoLogger,
+  ) {
+    this.logger.setContext(DataExportService.name);
+  }
+
+  /** Enqueue a data-export job for `user`, rate-limited per account. The client
+   * polls `GET /jobs/:id` and downloads `GET /jobs/:id/result` once it
+   * succeeds; the link stops working after the job's `expiresAt`. */
+  async request(user: User): Promise<ConversionJobView> {
+    const limit = await this.rateLimits.hit(
+      'data_export',
+      `user:${user.id}`,
+      MAX_REQUESTS_PER_WINDOW,
+      WINDOW_SECONDS,
+    );
+    if (!limit.allowed) {
+      throw new HttpException(
+        { code: 'rate_limited', details: { retryAfterSeconds: limit.retryAfterSeconds } },
+        HttpStatus.TOO_MANY_REQUESTS,
+      );
+    }
+
+    const job = await this.jobs.enqueue({
+      ownerId: user.id,
+      kind: DATA_EXPORT_KIND,
+      from: 'account',
+      to: 'zip',
+      // A data export gathers everything from the database at build time, so the
+      // job needs no input payload.
+      input: Buffer.alloc(0),
+    });
+    this.logger.info({ jobId: job.id, userId: user.id }, 'audit: data export requested');
+    return this.jobs.viewOf(job);
+  }
+
+  /** Worker entry point (via {@link DATA_EXPORT_PROCESSOR}): build the ZIP for a
+   * requested export. */
+  async build(job: ConversionJob): Promise<{ bytes: Buffer; mimeType: string }> {
+    const user = await this.prisma.user.findUniqueOrThrow({ where: { id: job.ownerId } });
+
+    const archive = archiver('zip', { zlib: { level: 9 } });
+    const chunks: Buffer[] = [];
+    archive.on('data', (chunk: Buffer) => chunks.push(chunk));
+    const finished = new Promise<void>((resolve, reject) => {
+      archive.on('end', () => resolve());
+      archive.on('error', reject);
+    });
+
+    archive.append(JSON.stringify(this.profileOf(user), null, 2), { name: 'profile.json' });
+    archive.append(JSON.stringify(await this.membershipsOf(user), null, 2), {
+      name: 'memberships.json',
+    });
+
+    // Only ponds this user owns — their personal pond and any shared ponds they
+    // created. Foreign ponds they are merely a member of are represented by the
+    // memberships list, not their content.
+    const ponds = await this.prisma.pond.findMany({
+      where: { ownerId: user.id, deletedAt: null },
+      select: { id: true, slug: true },
+      orderBy: { slug: 'asc' },
+    });
+    for (const pond of ponds) {
+      await this.exports.appendPondMarkdown(archive, user, pond, `ponds/${pond.slug}/`);
+    }
+
+    await archive.finalize();
+    await finished;
+    this.logger.info(
+      { jobId: job.id, userId: user.id, ponds: ponds.length },
+      'audit: data export built',
+    );
+    return { bytes: Buffer.concat(chunks), mimeType: 'application/zip' };
+  }
+
+  /** Drop the stored bytes of every export whose link has expired (data
+   * minimization: the archive is transient). The row remains for its audit
+   * trail; only the payload is deleted, so a later download 404s. Registered as
+   * a scheduled maintenance job. */
+  async purgeExpired(): Promise<number> {
+    const result = await this.prisma.conversionJob.updateMany({
+      where: { kind: DATA_EXPORT_KIND, expiresAt: { lt: new Date() }, result: { not: null } },
+      data: { result: null, resultMimeType: null },
+    });
+    if (result.count > 0) {
+      this.logger.info({ purged: result.count }, 'audit: expired data exports purged');
+    }
+    return result.count;
+  }
+
+  /** The requester's own account fields — the data minimization set the app
+   * collects (security.md §Privacy): identity, locale, and lifecycle stamps. */
+  private profileOf(user: User): Record<string, unknown> {
+    return {
+      id: user.id,
+      username: user.username,
+      email: user.email,
+      displayName: user.displayName,
+      locale: user.locale,
+      isSiteAdmin: user.isSiteAdmin,
+      status: user.status,
+      emailVerifiedAt: user.emailVerifiedAt?.toISOString() ?? null,
+      createdAt: user.createdAt.toISOString(),
+      lastLoginAt: user.lastLoginAt?.toISOString() ?? null,
+    };
+  }
+
+  /** Every pond the user has a role grant in, with the grants that concern
+   * them. This is the user's own membership data — pond identity plus their
+   * roles — and deliberately carries none of those ponds' content. */
+  private async membershipsOf(user: User): Promise<unknown[]> {
+    const grants = await this.prisma.roleGrant.findMany({
+      where: { subjectType: 'USER', subjectId: user.id },
+      include: { pond: { select: { slug: true, name: true, type: true, ownerId: true } } },
+      orderBy: { createdAt: 'asc' },
+    });
+    const byPond = new Map<
+      string,
+      { pondSlug: string; pondName: string; pondType: string; owner: boolean; grants: unknown[] }
+    >();
+    for (const grant of grants) {
+      const entry = byPond.get(grant.pondId) ?? {
+        pondSlug: grant.pond.slug,
+        pondName: grant.pond.name,
+        pondType: grant.pond.type,
+        owner: grant.pond.ownerId === user.id,
+        grants: [],
+      };
+      entry.grants.push({
+        role: grant.role,
+        scopeType: grant.scopeType,
+        scopeId: grant.scopeId,
+        effect: grant.effect,
+        createdAt: grant.createdAt.toISOString(),
+      });
+      byPond.set(grant.pondId, entry);
+    }
+    return [...byPond.values()];
+  }
+}

apps/api/src/import-export/export.service.ts

@@ -59,6 +59,34 @@ export class ExportService {
     const pond = await this.prisma.pond.findFirst({ where: { id: pondId, deletedAt: null } });
     if (!pond) throw new NotFoundException();
 
+    const archive = archiver('zip', { zlib: { level: 9 } });
+    res.set('Content-Type', 'application/zip');
+    res.set('Content-Disposition', `attachment; filename="${pond.slug}.zip"`);
+    res.set('X-Content-Type-Options', 'nosniff');
+    archive.on('error', (error) => {
+      this.logger.error({ pondId, err: error.message }, 'pond export archive failed');
+      res.destroy(error);
+    });
+    archive.pipe(res);
+    await this.appendPondMarkdown(archive, user, pond);
+    await archive.finalize();
+  }
+
+  /**
+   * Append one pond's readable pages (as Markdown) and their media to an open
+   * archive, each entry under `prefix`. Shared by the pond ZIP above and the
+   * account data export (#68), which nests several ponds under `ponds/<slug>/`
+   * — the read filter runs here, so neither caller can leak a page the
+   * requester may not read. Page Markdown is appended as small strings; media
+   * as read streams, keeping memory bounded regardless of pond size.
+   */
+  async appendPondMarkdown(
+    archive: archiver.Archiver,
+    user: User,
+    pond: { id: string; slug: string },
+    prefix = '',
+  ): Promise<void> {
+    const pondId = pond.id;
     const pages = await this.prisma.page.findMany({
       where: { pondId, deletedAt: null },
       orderBy: { title: 'asc' },
@@ -99,16 +127,6 @@ export class ExportService {
       attachments.map((a) => [a.id, `${a.id}.${imageExtension(a.mimeType)}`]),
     );
 
-    const archive = archiver('zip', { zlib: { level: 9 } });
-    res.set('Content-Type', 'application/zip');
-    res.set('Content-Disposition', `attachment; filename="${pond.slug}.zip"`);
-    res.set('X-Content-Type-Options', 'nosniff');
-    archive.on('error', (error) => {
-      this.logger.error({ pondId, err: error.message }, 'pond export archive failed');
-      res.destroy(error);
-    });
-    archive.pipe(res);
-
     for (const page of readablePages) {
       const markdown = markdownForZip(
         page.contentCache?.markdown ?? '',
@@ -116,7 +134,7 @@ export class ExportService {
         mediaNameById,
       );
       // Page slugs are unique within a pond, so `<slug>.md` never collides.
-      archive.append(markdown, { name: `${page.slug}.md` });
+      archive.append(markdown, { name: `${prefix}${page.slug}.md` });
     }
     for (const attachment of attachments) {
       const stream = this.storage.createReadStream(pond.id, attachment.id);
@@ -128,13 +146,12 @@ export class ExportService {
           'export: media read failed',
         ),
       );
-      archive.append(stream, { name: `media/${mediaNameById.get(attachment.id)!}` });
+      archive.append(stream, { name: `${prefix}media/${mediaNameById.get(attachment.id)!}` });
     }
     this.logger.info(
       { pondId, pages: readablePages.length, media: attachments.length, userId: user.id },
       'audit: pond exported as markdown zip',
     );
-    await archive.finalize();
   }
 
   /**

apps/api/src/import-export/import-export.module.ts

@@ -1,10 +1,15 @@
-import { Module } from '@nestjs/common';
+import { Module, OnModuleInit } from '@nestjs/common';
 
 import { FilesModule } from '../files/files.module';
 import { PagesModule } from '../pages/pages.module';
+import { SchedulerModule } from '../scheduler/scheduler.module';
+import { SchedulerService } from '../scheduler/scheduler.service';
 
 import { ConversionJobService } from './conversion-job.service';
 import { ConversionWorker } from './conversion-worker.service';
+import { DATA_EXPORT_PROCESSOR } from './data-export.constants';
+import { DataExportController } from './data-export.controller';
+import { DataExportService } from './data-export.service';
 import { ExportController } from './export.controller';
 import { ExportService } from './export.service';
 import { GotenbergHttpRenderer, GotenbergRenderer } from './gotenberg.renderer';
@@ -14,23 +19,29 @@ import { ImportService } from './import.service';
 import { JobsController } from './jobs.controller';
 import { PandocConverter, PandocServerConverter } from './pandoc.converter';
 
+/** How often expired data-export payloads are purged (#68). Hourly is ample:
+ * the link's own expiry check already stops downloads the moment it lapses. */
+const EXPORT_PURGE_CADENCE_SECONDS = 60 * 60;
+
 /**
  * Import/export orchestration (ADR 0009): the conversion job queue, its worker,
- * the pandoc-server client (#62), and the document import pipeline (#63, which
- * turns an uploaded `.docx`/`.odt` into a new page). Export (#65) will add its
- * feature endpoints here too.
+ * the pandoc-server client (#62), the document import pipeline (#63), the
+ * feature exports (#65/#67), and the GDPR account data export (#68).
  */
 @Module({
-  imports: [FilesModule, PagesModule],
-  controllers: [JobsController, ImportController, ExportController],
+  imports: [FilesModule, PagesModule, SchedulerModule],
+  controllers: [JobsController, ImportController, ExportController, DataExportController],
   providers: [
     ConversionJobService,
     ConversionWorker,
     ImportService,
     ExportService,
+    DataExportService,
     // The worker resolves the import pipeline through this token (never the
     // class), so its file does not import the import service's (avoids a cycle).
     { provide: IMPORT_PROCESSOR, useExisting: ImportService },
+    // Same token indirection for the data-export builder (#68).
+    { provide: DATA_EXPORT_PROCESSOR, useExisting: DataExportService },
     // Bind the abstract converter to the HTTP implementation; tests override
     // this provider with a fake so the queue mechanics need no live sidecar.
     { provide: PandocConverter, useClass: PandocServerConverter },
@@ -39,4 +50,17 @@ import { PandocConverter, PandocServerConverter } from './pandoc.converter';
   ],
   exports: [ConversionJobService, PandocConverter],
 })
-export class ImportExportModule {}
+export class ImportExportModule implements OnModuleInit {
+  constructor(
+    private readonly scheduler: SchedulerService,
+    private readonly dataExport: DataExportService,
+  ) {}
+
+  onModuleInit(): void {
+    this.scheduler.register({
+      name: 'data-export-purge',
+      cadenceSeconds: EXPORT_PURGE_CADENCE_SECONDS,
+      run: () => this.dataExport.purgeExpired().then(() => undefined),
+    });
+  }
+}

apps/api/src/import-export/import.service.ts

@@ -171,6 +171,7 @@ export class ImportService implements ImportProcessor {
         targetFormat: 'page',
         errorCode: null,
         resultPageId: page.id,
+        expiresAt: null,
         createdAt: now,
         updatedAt: now,
       };

apps/web/src/export/job-download.ts

@@ -0,0 +1,46 @@
+import type { ConversionJobView } from '@dorfteich/shared';
+
+import { ApiError, apiGet } from '../lib/api';
+
+const POLL_INTERVAL_MS = 1000;
+/** Default poll budget — a whole-account export (#68) can take longer than a
+ * single-page conversion, so callers may raise or lower it. */
+const DEFAULT_MAX_POLLS = 300;
+
+const delay = (ms: number): Promise<void> => new Promise((resolve) => setTimeout(resolve, ms));
+
+/** Poll a conversion job until it succeeds or fails, or the poll budget runs
+ * out (returns whatever the last status was). Shared by every export hook. */
+export async function pollJob(
+  job: ConversionJobView,
+  maxPolls = DEFAULT_MAX_POLLS,
+): Promise<ConversionJobView> {
+  let current = job;
+  for (
+    let poll = 0;
+    current.status !== 'succeeded' && current.status !== 'failed' && poll < maxPolls;
+    poll += 1
+  ) {
+    await delay(POLL_INTERVAL_MS);
+    current = await apiGet<ConversionJobView>(`/jobs/${current.id}`);
+  }
+  return current;
+}
+
+/** Download a finished job's result and save it as `fileName`. The result
+ * endpoint is session-authenticated, so a plain `fetch` carries the cookie. */
+export async function downloadJobResult(jobId: string, fileName: string): Promise<void> {
+  const response = await fetch(`/api/v1/jobs/${jobId}/result`);
+  if (!response.ok) {
+    throw new ApiError(response.status, { code: `http_${response.status}`, message: '' });
+  }
+  const blob = await response.blob();
+  const url = URL.createObjectURL(blob);
+  const anchor = document.createElement('a');
+  anchor.href = url;
+  anchor.download = fileName;
+  document.body.appendChild(anchor);
+  anchor.click();
+  anchor.remove();
+  URL.revokeObjectURL(url);
+}

apps/web/src/export/use-data-export.ts

@@ -0,0 +1,54 @@
+import type { ConversionJobView } from '@dorfteich/shared';
+import { useCallback, useState } from 'react';
+
+import { ApiError, apiPost } from '../lib/api';
+
+import { downloadJobResult, pollJob } from './job-download';
+
+export type DataExportStatus = 'idle' | 'busy' | 'ready' | 'error' | 'rateLimited';
+
+export interface UseDataExport {
+  status: DataExportStatus;
+  /** When the ready download link expires (ISO), or null. */
+  expiresAt: string | null;
+  request: () => void;
+  download: () => void;
+}
+
+/**
+ * Self-service GDPR data export (issue #68): request the export, poll the job,
+ * and — once it is ready — download the ZIP. A recent request is rejected with
+ * 429, surfaced as a distinct `rateLimited` status.
+ */
+export function useDataExport(): UseDataExport {
+  const [status, setStatus] = useState<DataExportStatus>('idle');
+  const [expiresAt, setExpiresAt] = useState<string | null>(null);
+  const [jobId, setJobId] = useState<string | null>(null);
+
+  const request = useCallback((): void => {
+    void (async () => {
+      setStatus('busy');
+      setExpiresAt(null);
+      setJobId(null);
+      try {
+        const job = await pollJob(await apiPost<ConversionJobView>('/users/me/data-export'));
+        if (job.status === 'succeeded') {
+          setJobId(job.id);
+          setExpiresAt(job.expiresAt);
+          setStatus('ready');
+        } else {
+          setStatus('error');
+        }
+      } catch (error) {
+        setStatus(error instanceof ApiError && error.status === 429 ? 'rateLimited' : 'error');
+      }
+    })();
+  }, []);
+
+  const download = useCallback((): void => {
+    if (!jobId) return;
+    void downloadJobResult(jobId, 'dorfteich-export.zip').catch(() => setStatus('error'));
+  }, [jobId]);
+
+  return { status, expiresAt, request, download };
+}

apps/web/src/export/use-document-export.ts

@@ -1,31 +1,16 @@
 import type { ConversionJobView, ExportFormat } from '@dorfteich/shared';
 import { useCallback, useState } from 'react';
 
-import { ApiError, apiGet, apiPost } from '../lib/api';
+import { apiPost } from '../lib/api';
+
+import { downloadJobResult, pollJob } from './job-download';
 
 export type ExportStatus = 'idle' | 'busy' | 'error';
 
-const POLL_INTERVAL_MS = 1000;
+/** A single-page conversion is quick — cap the poll budget below the shared
+ * default. */
 const MAX_POLLS = 180;
 
-const delay = (ms: number): Promise<void> => new Promise((resolve) => setTimeout(resolve, ms));
-
-/** Fetch a finished job's result and save it under `fileName`. The result
- * endpoint is session-authenticated, so a plain `fetch` carries the cookie. */
-async function downloadResult(jobId: string, fileName: string): Promise<void> {
-  const response = await fetch(`/api/v1/jobs/${jobId}/result`);
-  if (!response.ok) throw new ApiError(response.status, { code: 'conversion_failed', message: '' });
-  const blob = await response.blob();
-  const url = URL.createObjectURL(blob);
-  const anchor = document.createElement('a');
-  anchor.href = url;
-  anchor.download = fileName;
-  document.body.appendChild(anchor);
-  anchor.click();
-  anchor.remove();
-  URL.revokeObjectURL(url);
-}
-
 export interface UseDocumentExport {
   /** Per-format status so each button shows its own progress/error. */
   status: Partial<Record<ExportFormat, ExportStatus>>;
@@ -48,17 +33,12 @@ export function useDocumentExport(): UseDocumentExport {
       void (async () => {
         set(format, 'busy');
         try {
-          let job = await apiPost<ConversionJobView>(`/pages/${pageId}/export`, { format });
-          for (
-            let poll = 0;
-            job.status !== 'succeeded' && job.status !== 'failed' && poll < MAX_POLLS;
-            poll += 1
-          ) {
-            await delay(POLL_INTERVAL_MS);
-            job = await apiGet<ConversionJobView>(`/jobs/${job.id}`);
-          }
+          const job = await pollJob(
+            await apiPost<ConversionJobView>(`/pages/${pageId}/export`, { format }),
+            MAX_POLLS,
+          );
           if (job.status === 'succeeded') {
-            await downloadResult(job.id, `${slug}.${format}`);
+            await downloadJobResult(job.id, `${slug}.${format}`);
             set(format, 'idle');
           } else {
             set(format, 'error');

apps/web/src/pages/SettingsPage.tsx

@@ -7,6 +7,7 @@ import { useTranslation } from 'react-i18next';
 
 import { useAuth } from '../auth/auth-context';
 import { Field, FormError, FormSuccess, applyFieldErrors } from '../components/forms';
+import { useDataExport } from '../export/use-data-export';
 import { apiDelete, apiGet, apiPatch, apiPost } from '../lib/api';
 
 interface SessionView {
@@ -25,10 +26,57 @@ export function SettingsPage(): React.JSX.Element {
       <ProfileSection />
       <PasswordSection />
       <SessionsSection />
+      <DataExportSection />
     </>
   );
 }
 
+function DataExportSection(): React.JSX.Element {
+  const { t, i18n } = useTranslation();
+  const { status, expiresAt, request, download } = useDataExport();
+
+  const formatTime = (iso: string): string =>
+    new Intl.DateTimeFormat(i18n.language, { dateStyle: 'medium', timeStyle: 'short' }).format(
+      new Date(iso),
+    );
+
+  return (
+    <section className="settings-section">
+      <h2>{t('settings:dataExport.title')}</h2>
+      <p className="sidebar__hint">{t('settings:dataExport.description')}</p>
+      {status === 'error' && (
+        <p className="form-banner form-banner--error" role="alert">
+          {t('settings:dataExport.failed')}
+        </p>
+      )}
+      {status === 'rateLimited' && (
+        <p className="form-banner form-banner--error" role="alert">
+          {t('settings:dataExport.rateLimited')}
+        </p>
+      )}
+      {status === 'ready' ? (
+        <>
+          <FormSuccess message={t('settings:dataExport.ready')} />
+          <button type="button" className="button" onClick={download}>
+            {t('settings:dataExport.download')}
+          </button>
+          {expiresAt && (
+            <p className="sidebar__hint">
+              {t('settings:dataExport.expiresHint', { when: formatTime(expiresAt) })}
+            </p>
+          )}
+        </>
+      ) : (
+        <button type="button" className="button" onClick={request} disabled={status === 'busy'}>
+          {status === 'busy'
+            ? t('settings:dataExport.preparing')
+            : t('settings:dataExport.request')}
+        </button>
+      )}
+    </section>
+  );
+}
+
 function ProfileSection(): React.JSX.Element {
   const { t, i18n } = useTranslation();
   const { user, refresh } = useAuth();

packages/shared/i18n/de/settings.json

@@ -25,6 +25,17 @@
     "revokeAll": "Alle anderen Sitzungen abmelden",
     "empty": "Keine weiteren aktiven Sitzungen."
   },
+  "dataExport": {
+    "title": "Meine Daten exportieren",
+    "description": "Lade ein ZIP mit deinem Profil, einer Liste deiner Mitgliedschaften und dem Markdown-Export deines persönlichen Teichs sowie der Teiche, die dir gehören.",
+    "request": "Export vorbereiten",
+    "preparing": "Dein Export wird vorbereitet…",
+    "download": "Export herunterladen",
+    "ready": "Dein Export ist bereit.",
+    "expiresHint": "Der Download-Link läuft am {{when}} ab.",
+    "failed": "Der Export konnte nicht erstellt werden. Bitte versuche es erneut.",
+    "rateLimited": "Du hast kürzlich einen Export angefordert. Bitte versuche es später erneut."
+  },
   "admin": {
     "title": "Administration",
     "instanceName": "Name der Instanz",

packages/shared/i18n/en/settings.json

@@ -25,6 +25,17 @@
     "revokeAll": "Sign out all other sessions",
     "empty": "No other active sessions."
   },
+  "dataExport": {
+    "title": "Export my data",
+    "description": "Download a ZIP with your profile, a list of your memberships, and the Markdown export of your personal pond and the ponds you own.",
+    "request": "Prepare export",
+    "preparing": "Preparing your export…",
+    "download": "Download export",
+    "ready": "Your export is ready.",
+    "expiresHint": "The download link expires on {{when}}.",
+    "failed": "The export could not be prepared. Please try again.",
+    "rateLimited": "You requested an export recently. Please try again later."
+  },
   "admin": {
     "title": "Administration",
     "instanceName": "Instance name",

packages/shared/src/conversion.ts

@@ -21,6 +21,10 @@ export interface ConversionJobView {
    * the client can navigate to it. `null` for a pending/failed import and for
    * plain byte→byte conversions (export). */
   resultPageId: string | null;
+  /** For a data-export job (#68), when its download link stops working — the
+   * result is purged after this. `null` for every other job kind, which never
+   * expires. */
+  expiresAt: string | null;
   createdAt: string;
   updatedAt: string;
 }