Rice

Welcome to Rice

Cognitive Substrate for AI Agents

Rice is a complete memory infrastructure that gives AI agents the ability to remember, learn, and adapt. High-performance storage meets cognitive memory architecture.

Get startedView on GitHub

Why Rice?

    Cognitive Memory Architecture
    Four-component memory system inspired by human cognition. Working, Episodic, Procedural, and Semantic memory for building truly intelligent agents.
    Working Memory
    Dynamic attention-based context management. Items decay naturally over time if not accessed, simulating real working memory for your agents.
    Episodic Memory
    Record and recall past experiences. Enable few-shot learning from history with semantic search over traces of Input, Action, Outcome, and Reasoning.
    Procedural Memory
    Execute compiled WebAssembly skills server-side. Deterministic, sandboxed logic for calculations, tools, and reusable procedures.
    Built for Agents, Not Chatbots
    Unlike vector databases or simple memory layers, Rice provides a complete cognitive substrate with execution, learning, and attention.