AddThisFeature

Personal Data Export

Give a user everything you hold about them, safely.

involved Privacy & Compliance

What it adds

A subject-access-request flow that gathers one person's data from every table and file store, packages it, and delivers it to a verified requester.

What your agent is told to do

6
  1. 1

    Enumerate every place personal data lives for a single user — profile, content, comments, activity, sessions, billing, support messages, uploaded files — and assemble the export from that list, not from whatever a list view happens to show.

  2. 2

    Verify identity before generating and again before releasing the file. Re-authenticate, and require the account's second factor if one is enabled.

  3. 3

    Generate asynchronously and notify the user when it is ready. A complete export can take minutes and must not run in a web request.

  4. 4

    Package machine-readable structured data alongside a plain human-readable index that explains what each file contains.

  5. 5

    Deliver via a single-use, short-lived, unguessable link, and delete the generated archive when it expires.

  6. 6

    Do NOT include other people's data. Shared threads, workspace records, and messages the user received contain third parties — include the requester's contributions and redact the rest, and never include another user's email or personal fields.

Edge cases it handles

7
  • The export archive is itself a breach vector: it is the most concentrated copy of a person's data you will ever produce. Encrypt at rest, expire fast, and never email it as an attachment.
  • Jointly owned workspace data belongs to the workspace, not the individual. Decide and document what a member gets versus what an owner gets.
  • Do not email the download link to an address the requester changed minutes earlier — check for a recent email change and require confirmation.
  • Rate-limit requests. Repeated export generation is an easy denial-of-service and an easy way to leave copies lying around.
  • Derived and inferred data counts too: computed scores, segments, and internal tags about the person belong in the export.
  • Very large accounts need chunked archives or streaming; a single multi-gigabyte file will fail on download.
  • Log that an export was requested, generated, and downloaded, including by whom — but do not log its contents.

Definition of done

9
  • The export covers every store holding the requester's personal data, per a documented data map.
  • Identity is verified before generation and before download.
  • Generation is asynchronous and the user is notified on completion.
  • The archive contains structured data plus a readable explanation of its contents.
  • Download links are single-use, expiring, and unguessable, and the archive is deleted on expiry.
  • No third party's personal data appears in the export.
  • Export requests are rate-limited and audit-logged.
  • The feature matches the existing design system.
  • No existing functionality is broken.

Related features

How it works

  1. 1

    Copy the link

    Grab the Markdown instruction URL for this feature.

  2. 2

    Give it to your AI

    Paste it into Claude Code, Cursor, v0, Lovable — whatever you build with.

  3. 3

    It inspects, then implements

    Your agent reads your existing app first, then adds the feature to fit it.

Works with your stack

These instructions are written to adapt. They tell the agent to detect your framework, match your existing design system, and reuse what you already have — rather than assuming a particular stack.

Need it tighter than that? Customize the feature and tell it exactly what you're running.