liteflow: a rules engine framework that orchestrates complex business logic and AI Agents as unified components

What it solves

LiteFlow is designed to decouple complex business logic and reduce code bloat in large systems. It allows developers to move hard-coded process flows into externalized rules, making complex orchestration—including synchronous and asynchronous mixing—easier to manage and modify without restarting the application.

How it works

The framework uses a Domain Specific Language (DSL) to define business rules in XML, JSON, or YML files. These rules drive the execution of "components" (small, unified logic units). It supports hot-refreshing rules from various sources like Nacos, Redis, or Zookeeper. Recently, it integrated a ReAct (Reasoning + Acting) Agent module, allowing AI Agents to be treated as standard components that can be orchestrated using the same logic operators (THEN, WHEN, IF, etc.) as traditional business nodes.

Who it’s for

It is built for Java developers (JDK 8 to 25) using Spring Boot or other Java frameworks who need to manage highly complex, componentized business orchestration.

Highlights

• AI Agent Integration: Wraps ReAct Agents as first-class components, enabling multi-agent collaboration and AI-driven decision making within business rules.
• Hot Deployment: Update business rules instantly without application restarts.
• Multi-Language Support: Supports scripting nodes in Groovy, Java, Kotlin, JavaScript, Python, Lua, and more.
• Flexible Orchestration: Supports complex layouts including parallel execution (WHEN), sequential flow (THEN), and conditional routing (IF).
• Broad Infrastructure Support: Rules can be loaded from databases, Nacos, Etcd, Zookeeper, Apollo, and Redis.