Leaving should be one click, not a support ticket

Export is the feature nobody uses and everybody should check before they start, because it is the only thing standing between “I chose this app” and “I am stuck in this app”. Years of tasks are a record of what you did, and a product that makes them awkward to retrieve has quietly made the decision to leave expensive.

The common failure is not a missing button. It is a button that is a paid feature, or one that produces something you cannot read, or a “request your data” form that arrives by email in 30 days. The test is whether you can get your data out right now, on the plan you are on, without asking anyone.

One JSON file, named for the day you took it, downloaded straight from the browser. On iPhone it goes to the share sheet. Inside:

• Projects — name, description, the area they belong to, their deadline, and when each was created and completed
• Sections within each project
• Tasks — title, notes, start date, deadline, the project and section they sit in, and their repeat rule
• Checklist items on each task, with their state
• Completed and cancelled tasks, with the dates they were completed or cancelled

It is indented and human-readable. You can open it in a text editor and understand it without a tool, which is the property that actually matters in ten years when the tool has gone.

There is no plan check on the export. It is not a Premium feature, not a retention lever, and not something that appears when you try to cancel.

One detail worth stating because it cuts the other way from how these things usually go: the free plan shows the last 14 days of completed tasks in the Logbook, but the export applies no such cutoff. A free account exports its whole completed history, including the part the app will not show it.

An archive, not a backup. Four specific losses, and it is better that you read them here than discover them later:

• Repeat rules become sentences. A rule is written out as English — every 2 weeks on Mon, every month on the last day — rather than as structured fields. Readable, but not something a machine can reconstruct exactly.
• An empty area vanishes. Areas are not exported as objects; an area is a name attached to a project. An area with no projects in it is not in the file, and area ordering is not either.
• Notes become plain text. Formatting and links in a task’s notes are flattened to their text.
• Nothing has an id. Relationships are expressed by name, so two projects with the same name are ambiguous in the file.

Trashed tasks are also excluded, which is intentional — the trash is already the place for things you deleted.

Export is one of the better-behaved corners of this market: the major apps all have one, and only Todoist's whole-account backup sits behind a paid plan. The differences are in the formats offered and in whether you can get the file immediately.

Every unknown above is a cell where the vendor does not document the answer plainly and this site will not guess. Apple is the largest gap — it publishes extensively about Reminders and almost nothing about getting data out of it — and Cultured Code document the route out without documenting what survives it.

Todoist is the useful comparison, and it splits: its CSV is a better interchange format than our JSON and it can import one, while exporting one project at a time without completed tasks is a worse archive than a single file with your whole history in it. Note neither product survives the round trip with repeats intact.

The place TaskPocket is clearly behind is the last row: there is no importer. Getting out is one click; getting in is copy and paste. If you have a large list somewhere else, that is the real cost of moving, and it is worth weighing before you start.