Introduction

Kairo (from Greek Kairos — the decisive moment for action) is a Java Agent operating system that provides a complete runtime environment for AI agents. Rather than being yet another LLM wrapper, Kairo models the agent runtime as an operating system, mapping every component to a familiar OS concept.

Kairo is not a wrapper — it's infrastructure. Think Netty for networking, Jackson for serialization, Kairo for AI Agents.

OS Concept Mapping

OS Concept	Kairo Mapping	Description	Status
Memory	Context	Context window as bounded memory with intelligent compaction	Implemented
System Call	Tool	56 built-in tools — the agent's interface to the outside world	Implemented
Process	Agent	Independent execution unit driven by a ReAct loop	Implemented
File System	Memory	Persistent knowledge storage (file / in-memory / JDBC)	Implemented
Signal	Hook	10 hook points with CONTINUE/MODIFY/SKIP/ABORT/INJECT decisions	Implemented
Executable	Skill	Plug-and-play capability packs in Markdown format	Implemented
Job Scheduling	Cron + Team	Scheduled tasks + multi-agent team orchestration	Implemented
IPC	A2A Protocol	Agent-to-Agent communication for cross-agent invocation	Implemented
Middleware	Middleware Pipeline	Declarative request/response interception	Implemented
Package Manager	Plugin	Install/enable/disable skills, hooks, MCP servers, agents (Claude Code format compatible)	Implemented
Network	Gateway	Multi-channel routing, session management, streaming, mirroring	Implemented
Device Driver	Channel	IM adapter layer (DingTalk, Feishu, Slack, Telegram planned)	Implemented
Cron	Cron	Scheduled task execution with flexible scheduling	Implemented
Checkpoint	Snapshot	Agent state serialization and restoration	Implemented

Why Kairo

Kairo is built on Project Reactor for fully reactive, non-blocking execution and supports Claude, GLM, Qwen, GPT, and other models out of the box. The framework is model-agnostic — swap providers without changing agent logic.