Lifecycle & history

Knowledge changes. A process gets rewritten, a duplicate topic is merged into another, an old document is archived. The API gives you the latest, trustworthy state by default, and lets you opt into the full history when you need it.

Latest by default

Unless you ask otherwise, the topic and item endpoints return only the current state of the knowledge base. Records that have been archived or superseded by a newer version are left out. Search always behaves this way and never returns historical records.

This is what you want almost all of the time: an AI tool answering from your processes should use the current process, not a corrected-away older one.

Opting into history

The topic and item endpoints accept includeHistory=true. With it, the response also includes archived and superseded records, and each record carries two extra fields:

• status tells you whether a record is active, archived, or superseded.
• supersededById points to the record that replaced this one, so you can walk a correction chain forward to the current version.

curl "https://api.mereon.ai/v1/topics/top_123?includeHistory=true" \
  -H "Authorization: Bearer $MEREON_TOKEN"

Use history deliberately

Historical records exist for audit and lineage, not for answering questions. If you are feeding results to an assistant, leave includeHistory off so it never quotes superseded knowledge as if it were current.

Why a record might be missing

If a topic or item you expected is not in a default response, the likely reasons are:

• It was archived or superseded, so it only appears with includeHistory=true.
• It is outside what the token’s owner can see (see Permissions).