Skip to main content

What is the conversation graph?

The conversation graph is the structure the Verse Engine uses to represent a conversation. Rather than treating a conversation as a flat list of messages, the Verse Engine maps it as a graph — a network of contributions and the relationships between them.

The underlying structure is a DAG: a Directed Acyclic Graph.

What a DAG means for conversations​

In a DAG:

  • Each node is an Utterance — a single contribution from a participant
  • Each edge is a directed relationship between Utterances — indicating that one contribution responds to, follows from, or references another
  • The graph is acyclic — relationships move forward through the conversation, without loops

This structure allows the Verse Engine to understand not just what was said, but how it relates to everything else that has been said.

Why this matters​

A linear message queue can only answer the question: what was the last thing said? A conversation graph can answer much richer questions:

  • Which thread is this Utterance part of?
  • Which earlier point is this response addressing?
  • Which participant has the most relevant perspective to contribute at this moment?
  • Where has a line of discussion been fully explored, and where is it still open?

This is what allows AwakeVerse conversations to behave like real discussions rather than turn-taking exchanges. Participants respond to the threads most relevant to their perspective — not simply to the most recent message.

The graph in multi-participant conversations​

The value of the conversation graph becomes most apparent in Dialogue, where multiple participants are contributing simultaneously. Without a graph structure, a multi-participant conversation would quickly become incoherent — participants losing track of which points have been addressed, which threads are still active, and whose perspective is most relevant next.

The conversation graph gives the Verse Engine the structural information it needs to coordinate multiple independent participants in a coherent exchange. It is the foundation that resonance-based speaker selection operates on.

The graph over time​

The conversation graph grows as the conversation develops. Every new Utterance adds a node; every relationship between Utterances adds an edge. The engine reads this growing structure to understand where the conversation is, what has been resolved, and what still needs attention.

Related: What is an Utterance? · What is resonance-based speaker selection? · How does the Verse Engine work?