Memory Currents
Help guides

Help guides

Edit Memories

Correct a Memory's words while keeping every earlier version.

Fix a mistake without starting over

A Memory might have a typo in it, a misspelled name, or the wrong time for an appointment. You can correct its text where it is. The Memory keeps its date, Thoughts, types, tags, Currents, and to-do outcomes, and Memory Currents keeps every earlier version of the text.

Edit a Memory

Choose Tools, the wrench button at the top right of the Memory's card, then Edit Memory. Or choose Edit in the reader after you choose Expand Memory on a longer Memory. Correct the text and choose Save. The corrected text appears straight away, with the note “Saved. Earlier versions are kept in Revision History.” To stop without saving, choose Cancel. If you started from Tools, that takes you back to the card; if you started from Edit in the reader, it returns to the text.

Only active Memories can be edited. An archived Memory shows Archived across the top of its card and doesn't offer Edit Memory until you unarchive it. If it was edited before you archived it, its Revision History is still there to read.

If the Memory changed after you started editing — for example, an AI tool corrected it in the meantime — your edit isn't saved. Memory Currents shows the Latest version and keeps your text in the editor, so you can check it against the latest version and save again.

Corrections, not rewrites

Editing is for fixing what a Memory says, not for turning it into a different Memory. Memory Currents compares your new text with the old, and an edit that changes too much isn't saved. An edit that would remove one of the Memory's Thoughts entirely counts as too much as well.

When that happens, nothing is saved. Save is no longer offered; instead you have two choices, and Cancel leaves the Memory as it was:

  • Save as new Memory keeps your text as a new Memory. The original stays as it was.
  • Save as new and archive this one keeps your text as a new Memory, then archives the original. The original is archived only once the new Memory is saved, so if saving fails, nothing changes. If the new Memory is saved but archiving fails, the original is still here, and you can choose Archive from its card's Tools.

An archived original disappears from Active views, but its words, Thoughts, tags, and connections are preserved. Choose Archived or All to find it, then choose Unarchive and confirm to return it to active use.

What happens to Thoughts

When you edit a Memory, its Thoughts follow your correction:

  • A Memory with a single Thought: the Thought takes the corrected text.
  • A Thought that uses the Memory's own words, where your edit touched it, is updated with the edit.
  • A Thought in its own words — such as one reworded to stand on its own, or one that came from an import — is reworded from your correction shortly afterward. Until then, it shows its earlier wording.
  • Thoughts your edit didn't touch stay as they are.

Thoughts keep their order, types, and to-do outcomes. Editing never adds or removes a Thought. The short title written for the Memory, and for any Thought whose text changed, is written again for the new text.

If suggestions for splitting the Memory were waiting for you, they are withdrawn, because they came from the earlier text; you can ask for new ones. Suggestions still being prepared are prepared again from the corrected text.

You can also edit a smaller Memory created by splitting. Its own text changes, but the Memory it came from doesn't, and neither do its Thoughts, which belong to that original and keep their wording. A Thought's text can't be edited on its own.

Every version is kept

Each edit adds a new version of the Memory's text; earlier versions are never changed or removed. Once a Memory has been edited, choose Revision History from its card's Tools or in the reader to see every version, newest first, each with its date and text. The first version, marked Original, is the text as it was before your first edit. Each later version shows how it was made: Edited in the app, or Edited through an AI connection, so you can tell which changes your AI tool made. Back to Memory returns to the text, or takes you back to the card if you opened Revision History from Tools. Revision History stays available after you archive a Memory.

Export includes each Memory's current text. Earlier versions aren't included.

Edit through an AI tool

If you've connected an AI tool, you can ask it to correct a Memory — for example, “fix the spelling of Anna's name in yesterday's note about Florence.”

Your AI tool names the exact passages to change, character for character, and what each one becomes. It can change up to 20 passages at once. It doesn't restate the rest of the Memory, so words it didn't name stay exactly as they were. The same limit applies as in the app: if the edit changes too much, nothing is saved, and your AI tool can offer to save the new text as a new Memory instead and then archive the original.

When your AI tool looks through your Memories, it sees the first 2,000 characters of each one. To correct a longer Memory, it reads that Memory in full first, so it can correct any part of it.

If your AI tool loses the answer to an edit and tries the same edit again, the repeat is refused when the first attempt was saved, because the Memory has already changed. Your AI tool can look at the Memory again to check.

Memories explains what a Memory keeps and how it's organized.