Tags and organisation

Tags are a lightweight way to group related memories. A memory can have any number of tags, and you can list all memories sharing a tag with a single tool call.

How tags work

Tags are plain strings attached to a memory at creation time (or updated later). There's no predefined list — you choose whatever makes sense for your use case.

remember(
  key="project/myapp/auth-approach",
  value="Using JWT with 1-hour expiry and refresh tokens stored in httpOnly cookies.",
  tags=["project", "myapp", "auth", "decision"]
)

You can then retrieve all memories in a group:

List all memories tagged "decision".

Suggested tag conventions

By topic

Tag memories with the domain they belong to:

project, auth, database, api, frontend, infra, security

By type

Tag memories with what kind of information they contain:

decision, convention, reference, todo, context, preference

By project or team

Tag memories with the project or person they relate to:

myapp, client-acme, team-alice, sprint-5

Combining approaches

Most useful memories have multiple tags — one for topic, one for type:

tags=["myapp", "auth", "decision"]
tags=["preferences", "code-style"]
tags=["project", "deadline", "reference"]

Searching vs. listing by tag

Approach	Best for
list_memories(tag="auth")	You know the exact tag and want all memories in that group
search_memories("authentication")	You want memories semantically related to auth, regardless of how they were tagged

Tips

• Tag at write time — it's easier to be consistent if you tag memories when you create them
• Use the plural for groups — decisions, references, preferences reads naturally when listing
• Don't over-tag — 2–4 tags per memory is usually enough
• Combine with a key hierarchy — tags describe what category a memory is in; the key describes which specific thing it is