Your notes now flag when they go stale

Your knowledge base grows. Some of it goes out of date. A note that was true last year can quietly become wrong, and your AI cannot tell the difference. It reads the note, trusts it, and hands you old information as if it were current.

Now every note carries a freshness signal.

• Active. Confirmed recently. Good to use.
• Stale. Nobody has confirmed it in a while, or its review date has passed.
• Retired. Archived. Kept for reference, not maintained.

Stale notes are never hidden. They still show up in search and in your lists. They just carry a clear flag that says "this might be out of date, check it before relying on it," so Claude and ChatGPT stop presenting an old note as current fact.

When a note is still accurate, mark it verified. One click. That resets the clock.

Also new: tell a folder what does not belong in it. Folders already carry a description and AI instructions. Now they can also say what to keep out, like "no meeting notes, those live in Inbox." Your AI reads it and files notes in the right place.

For Claude and ChatGPT (MCP): every note now returns a freshness object (active, stale, or retired) with a short hint. Stale notes stay in results, just flagged. Your AI can mark a note verified once you confirm it, set a review date so a time-boxed note flags itself, and list stale notes when it is time to clean up. Folders return what does not belong in them too.

Live for everyone. Nothing to turn on.