This is a super basic microservice that just listens for speech and transcribes what it hears into the nexus with timestamps. Future versions should include speaker identity recognition (voice pitch/timbre, etc). For now we can start with basic ASR like Whisper or something.

This provides semantic embeddings for semantic search purposes

This is RAVEN's inner monologue. Basically, this is RAVEN thinking about what's going on, what he should do, etc. Thought should come before action. This is basically real-time thought.

The Dream Sequence is an anthropomorphic way of saying offline memory consolidation (OMC?). Basically, RAVEN will review memories to extract principles, lessons, and metaphors. Essentially, this is a way of simplifying and distilling memories so that they are easier to access and denser.

This is RAVEN's "executive" function - deciding when to act and what to say. Basically, this is a TTS microservice with a bit more.