Rice provides a complete cognitive infrastructure for AI agents. It consists of two main components:
| Component | Purpose |
|---|---|
| Storage | Storage and memory engine |
| Slate | State and runtime context management |
Rice's storage layer is a high-performance database designed for Multi-Agent AI Systems. It handles:
RiceDB uses Hyperdimensional Computing (HDC) for noise-tolerant, high-speed operations on memory traces.
Slate is the state and runtime context management layer that sits between your AI agents and storage. It provides a four-component cognitive memory architecture:
| Memory Type | Purpose |
|---|---|
| Working Memory | Pre-fetches context with decay and attention scoring |
| Episodic Memory | Stores interaction traces for learning from experience |
| Procedural Memory | Executes compiled skills server-side |
| Semantic Memory | Stores invariant facts and knowledge |
Slate turns stateless LLMs into learning, stateful agents.
Your AI Agent
↓
Slate (Memory Management)
↓
Storage (RiceDB)
focus, drift, commit, reminisce)This separation means you can focus on building intelligent agent behavior while Rice handles the memory infrastructure.