Semantic memory
Semantic memory is the layer of an AI agent that holds stable facts: things that are true about the user or the world that do not have a specific timestamp. 'Your wife is allergic to peanuts.' 'You prefer morning meetings.' Queries are by attribute, not time.
Why it matters
Episodic memory is "what happened." Semantic memory is "what is true." Without semantic, the agent has to re-derive every preference from the timeline on every query, which is slow and unreliable. With semantic, the agent can pull "user prefers morning meetings" directly without searching the prior 90 days of episodic notes.
Most semantic facts are not declared by the user; they are inferred from repeated episodic observations. The agent watched you book Felipe's three Saturdays in a row; it consolidated that into "default haircut is Felipe's on Saturday." Consolidation is the bridge between the layers.
How Ajeris handles this
Ajeris semantic memory is a structured per-user store keyed by attribute (preference, fact, routine, contact, instruction). It is populated by the dreaming worker consolidation pass plus by explicit user "save this" commands. The retrieval layer at prompt-build time pulls relevant attributes into the agent's context without searching episodic notes.
