Importing from other tools

Files come in through one place: Settings→Data, under Import data. The Import buttons on Notes and on Boards open it, and so do "Import notes" and "Import a board" in the command palette. Recipes have their own door on the Recipes page. Everything lands in the workspace you are in when you import.

What it reads#

The importer tells the sources apart by the shape of what you drop, before it opens anything for real: a vault from a Notion export, a Todoist backup from a Keep Takeout, a Trello board from a Margin file. You never have to say which is which. Under the drop zone, Moving from another app? has a tile for Todoist, Trello and Google Keep that says how to get the export out.

How an import goes

1. Drop the file in, or pick it.
2. Read the preview: counts per kind, sample rows, and for a CSV, where it goes and Match your columns so your headers land on the right fields.
3. Leave Skip duplicates on unless you want copies. Nothing already there is replaced.
4. Import, and read the result: what was created and what was skipped.

A bank statement CSV imported as expenses gets more help: the column layout is remembered for that bank's next statement, transfers between your own accounts are kept out of spending, and rows you already imported are skipped. See Expenses and money.

Obsidian and Notion, in more detail#

Obsidian. The whole folder tree comes across as real folders. Note identity is the note's path, so importing the same vault twice merges into what is already there. That means you can bring a big vault over in stages.

Notion. Notion stamps a 32-character id onto every page and folder name. The importer strips them, rebuilds the page hierarchy as folders, and sends each database through the CSV route as a board by default. Where Notion writes a database twice, once as the current view and once as all rows, the full one wins.

Todoist, Trello and Google Keep#

Todoist. In Todoist, open Settings, then Backups, and download the latest backup. Drop the whole zip here and each project becomes a board; one project's CSV works too. Sections become columns (a project without sections gets To Do, In Progress and Done), sub-tasks become a checklist on their card, comments go into the card's description, and priorities p1 to p3 become urgent, high and normal. A due date with no time stays on that day wherever you are. Repeating tasks keep repeating when the schedule is a common one ("every day", "every 2 weeks", "every monday", "every 15th"). Anything it cannot read keeps its Todoist wording in the description, so you can set it again.

Trello. On the board, open the menu, then Print, export and share, then Export as JSON, and drop that file here. Lists become columns, and cards keep their description, due date, labels, checklists and comments. A card here has one checklist, so several Trello checklists are merged, each item named after its list. Archived lists and cards stay behind, and the preview counts them. People on cards and attachments do not come across.

Google Keep. In Google Takeout, select only Keep and download the archive. Drop the whole zip here. Notes keep their text, checklists, labels (as tags), pins and the dates they were made and last edited. Archived notes go into a folder called Archived, and notes in Keep's trash are left out. Pictures and recordings stay in your Takeout folder; the preview counts them.

Exporting again later. You can drop a newer export of the same thing whenever you like. A Todoist project or a Trello board that is already here (same name) is skipped whole, and a Keep note is matched by its title and the moment it was made, so only notes you wrote since the last export come in. Two Keep notes with the same title, made on the same day, are still two notes. The preview's button counts only what is new, and the same file dropped into a different workspace imports in full there.

Plan limits and big imports#

Recipes work the same way: that import saves what fits and tells you how many did not.

Recipes, from a link, a file or a paste#

On Recipes, the menu beside the search box has Import from a link or file, with three tabs: Find online, From a link, and Paste or upload.

• From a link: the page is read on the server. A site that publishes proper structured recipe data is parsed exactly; a page that does not is read by the assistant and marked for you to check.
• Paste or upload: Markdown, plain text, or a Margin recipe file.
• Fill-in templates: the same menu has Template (JSON) and Template (markdown), if you would rather type a batch.

Every route shows a review screen before anything is saved. If the batch is bigger than your plan holds (25 recipes on Free), it saves what fits and tells you how many did not. See Recipes and the meal plan.

After an import#

Imported content is picked up by the Mind like anything you write, in the background. The result screen offers Organize into my Mind to run a weave pass over it straight away, which spends weave credits. A big import is the one time the review queue in Organize fills up fast: collapse every group and use Accept all where a group is plainly right.

Your memory, from another Margin or another assistant#

What the Mind knows comes in through its own door, not through Import data.

• A Margin brain file (the JSON from Settings→Your Brain, yours or a memory pack someone shared with you): upload it with Choose a .json brain on the same page, read the counts of facts, patterns and routines, then press Import this brain. Near-matches update what is already there, so importing the same file twice adds nothing.
• What another assistant remembers about you: Mind→Bring your memory gives you a prompt to paste into that assistant, and you review its answer fact by fact before anything is kept.

Both are covered in The Mind.

A board can also stay connected to a Google Sheet in two-way sync, which is different from importing it once. See Integrations.

Getting it all back out again (Markdown for notes, JSON for the workspace and your recipes) is covered in Your data.