Skip to main content

Use tags without over-organizing

Tags in SecondLoop are lightweight. Use them when they improve retrieval later, not as required metadata for every note.

Add tags when a message deserves structure​

  • Open a message menu and choose Tags.
  • Select an existing tag or type a new one.
  • Save when you are done.

SecondLoop also includes built-in system tags for common areas such as Work, Personal, Family, Health, Finance, Study, Travel, Social, Home, and Hobby.

Use AI suggestions as a shortcut​

  • SecondLoop can infer a small set of tags from message text and attachment understanding.
  • Clear matches may be applied automatically.
  • Lower-confidence suggestions stay reviewable in the tag picker before you keep them.

To avoid tag sprawl, suggestions prefer existing tags when a close match already exists.

Filter a conversation with include and exclude tags​

  • Open the tag filter in chat.
  • First tap includes a tag.
  • Second tap excludes it.
  • Third tap clears it.

This is useful when you want to hide one topic, focus on a project, or quickly review only work / finance / travel items in the current conversation.

Ask AI inside a tag scope​

When a tag filter is active, Ask AI uses that narrower scope for retrieval in the current conversation.
That makes it easier to ask questions like:

  • “What are the open work tasks from this week?”
  • “Summarize finance notes without the travel items.”
  • “Draft a follow-up using only family-related notes.”

Keep tags clean over time​

  • If you like plain text, you can keep manual #tags on a line that starts with #, such as #work #finance.
  • SecondLoop syncs those manual #tags with message tags.
  • Deleting a custom tag removes it from tagged messages and strips matching manual #tags from message text.
  • If two tags mean the same thing, merge suggestions help consolidate them.

A simple tagging rule​

Use as few tags as you can get away with.

A good starting point is a short set of broad tags, then add more only when you repeatedly need a sharper filter. The goal is faster recall later, not a perfect taxonomy today.