Skip to content

Import and export

Find import and export in Settings › Import & export, or in the menu bar’s File › Import and File › Export. Imports go into the open workspace.

Import tasks from a CSV or JSON file, and export a workspace’s tasks as either.

One row per task, with a header row. Only title is required; headers match ignoring case, spaces and underscores (Due Date and due_date are the same column), and other columns are ignored, so most spreadsheets import as they are.

Column Values
title Required
description Markdown
status not_started, in_progress, completed, cancelled
priority low, medium, high
start_date YYYY-MM-DD
due_date YYYY-MM-DD
project A project’s title
pull_request_url A link to the task’s pull request
id, parent_id Make a row a subtask of the row whose id it names

The status column also accepts todo, done and canceled.

An export writes this shape, and an import also takes a bare array of tasks. Every key but title is optional, and subtasks can nest.

{
"format": "devbrain.tasks",
"version": 1,
"tasks": [
{
"title": "Launch v2",
"status": "in_progress",
"priority": "high",
"dueDate": "2026-11-05",
"project": "Website",
"subtasks": [{ "title": "Write copy" }]
}
]
}
  • Projects are matched by title, ignoring case. A project is never created: create it first. A task whose project isn’t found (or matches more than one) is imported without one, and the summary says so.
  • Empty fields get your defaults: no due date takes the Default due date (a subtask takes its parent’s), no priority the Default priority, and no status is Not started.
  • Rows with an error are skipped, and the rest are imported: an unknown status or priority, a date that isn’t YYYY-MM-DD, a pull request link that isn’t a web address, a missing title, or a parent_id that loops. The summary lists every skipped row by number, so you can fix them and import that file again.
  • Nothing is imported twice. A task whose id is already in the workspace is skipped, so importing the same export again adds nothing.
  • Links to events and notes, and when a task was created or completed, aren’t imported. A task imported as completed is completed as of the import.

A whole file is refused, with nothing imported, if it isn’t CSV or JSON, isn’t UTF-8 text, has no title column, or is bigger than 10 MB or 10,000 tasks.

Export tasks saves every task you made in DevBrain as CSV or JSON, with its subtasks. Tasks in the trash and tasks synced from Linear are left out. A CSV export is marked as UTF-8, so spreadsheet apps read its accents and emoji correctly.

Import notes from Markdown files (.md, .markdown or .txt), one note per file. Front matter is optional:

---
title: Kickoff
project: Website
---
# Agenda

Without a title, a note is named after its file. project files it in a project, matched by title as for tasks. A file that can’t be read is listed in the summary, and the others are still imported. Each import takes up to 1,000 files, of up to 5 MB each.

Images and links in imported notes are never downloaded; see what the editor doesn’t show.

Export notes asks for a folder, then writes every note (except those in the trash) as a Markdown file named after its title, with this front matter, into a new folder named DevBrain notes <date> inside it. Nothing already there is ever overwritten: names that clash get “ (2)“, “ (3)” and so on.