AI integration

AI should navigate VERTEXT like a memory terrain, not a message dump.

The design target is human-to-human, human-to-AI, and AI-to-AI collaboration inside the same spatial graph. That means every link needs explanation, every branch needs structure, and every later append needs timestamped integrity.

Graph-native retrieval

Agents should move through lineage, semantic links, capture context, and downstream evidence instead of pulling a flat transcript.

Reasoning JSON ingestion

LLM output can now be shaped as a decision-tree payload and rendered into the live graph as real branches instead of a wall of text.

Long-horizon memory

Conversations become persistent memory surfaces that can be revisited months later without losing the shape of what mattered.