Open-source durable execution engine for orchestrating microservices, AI agents and adaptive workflow graphs, with workers in multiple languages.
dromara/liteflow
A Java rule and component orchestration engine that defines business flows in a DSL, with hot rule refresh and AI Agent nodes.
- Stars
- 3,885
- Forks
- 526
- Open issues
- 65
- Last commit
- last month
What is it?
LiteFlow is a Java rule engine that treats all business logic as components and arranges them with a small DSL written in XML, JSON or YAML. Sequential, parallel, conditional and loop orchestration are expressed directly in the rule file, so complex business flows can be changed without touching the code that calls them.
Rules can be loaded from databases, Nacos, Etcd, Zookeeper, Apollo or Redis, and changes take effect without restarting the application. Components can be written in Java or in scripting languages including Groovy, Kotlin, JavaScript, Python, Lua and QLExpress, and they can be retried with per-component configuration. The framework works with Spring Boot 2.x through 4.x and with plain Spring, and runs on JDK 8 through 25.
Since v2.16.0, the liteflow-agent module wraps an LLM Agent as a standard component, so AI nodes can be placed alongside business nodes in the same rules.
Key features
- 01Business logic is organized as components wired together by a DSL in XML, JSON or YAML.
- 02Synchronous, asynchronous, parallel, conditional and nested orchestration are defined in rule files.
- 03Rules are hot-refreshed from databases or config centers such as Nacos, Etcd and Apollo.
- 04Components can be written in Java or in several scripting languages.
- 05AI Agents can be orchestrated as components alongside business nodes.
Use cases
- Decoupling a large Java system with complex, frequently changing business flows.
- Letting operations or product staff adjust process orchestration through rule files.
- Combining LLM Agent calls with existing business steps in one workflow.
Things to know
- The AI Agent modules are compiled and verified only on JDK 17 with AgentScope Java 2.0.2.
- The guarded local filesystem backend is not a security sandbox, so untrusted code needs a hardened Docker or remote backend.
Topics
FAQ
- What is liteflow?
- A Java rule and component orchestration engine that defines business flows in a DSL, with hot rule refresh and AI Agent nodes. LiteFlow is a Java rule engine that treats all business logic as components and arranges them with a small DSL written in XML, JSON or YAML. Sequential, parallel, conditional and loop orchestration are expressed directly in the rule file, so complex business flows can be changed without touching the code that calls them.
- Is liteflow free to use? What license does it have?
- liteflow is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
- Is liteflow still maintained?
- Updated within 6 months. The last commit to liteflow was on Sep 21, 2026.
- What can I use liteflow for?
- Decoupling a large Java system with complex, frequently changing business flows. Letting operations or product staff adjust process orchestration through rule files. Combining LLM Agent calls with existing business steps in one workflow.
- Where can I get liteflow?
- The source code and installation instructions are on GitHub: https://github.com/dromara/liteflow
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Java and Vue3 low-code platform that generates full CRUD systems from natural language, with built-in AI chat, knowledge base and workflow orchestration.
A Java framework for building agentic, workflow and multi-agent AI applications on top of Spring AI.
Open-source AI agent application platform built on Java, LangChain4j, Spring Boot and Vue 3, covering models, agents, RAG, Skills, MCP and text-to-SQL analysis.
Java implementation of Apache Dubbo, an RPC and microservice framework for building service communication, discovery and traffic management.
Annotation-driven Java admin framework: annotate a JPA entity with @Erupt and get a generated admin UI, RBAC, audit logs and REST APIs at runtime.