Agents-Flex

Agents-Flex is an elegant, lightweight, and powerful open-source framework developed in Java, designed to streamline the creation of applications based on Large Language Models (LLMs). Drawing inspiration from established frameworks like LangChain and LlamaIndex, Agents-Flex offers a developer-friendly, portable, and orchestratable API that is not tied to any specific Java framework, making it a highly versatile tool for modern AI development.

The core philosophy behind Agents-Flex is to provide simplicity without sacrificing power. It enables developers to build complex AI systems, from simple chatbots to sophisticated autonomous agents, with minimal boilerplate code. The framework is built to be highly extensible, allowing developers to easily integrate custom models, data sources, and tools to fit their specific needs. It has strong support for both internationally recognized models like OpenAI and a wide array of Chinese domestic models, making it a globally relevant solution.

How to use Agents-Flex

Getting started with Agents-Flex is straightforward for any Java developer. The process typically involves these steps:

1. Add Dependency: Integrate Agents-Flex into your project by adding it as a dependency in your `pom.xml` (for Maven) or `build.gradle` file. For Spring Boot users, a dedicated `spring-boot-starter` is available for seamless auto-configuration.
2. Instantiate an LLM: Configure and create an instance of your desired LLM. For example, connecting to OpenAI is as simple as a single line of code: `OpenAILlm llm = OpenAILlm.of("YOUR_API_KEY");`
3. Perform a Chat Completion: Interact with the model by calling the `chat` method: `String response = llm.chat("what is your name?");`
4. Implement Function Calling: To connect the LLM to your own tools, define a Java class with methods annotated with `@FunctionDef` and `@FunctionParam`. Agents-Flex handles the complex logic of generating the tool schema, sending it to the LLM, and parsing the response to invoke the correct local method.
5. Build a RAG Pipeline: For Retrieval-Augmented Generation, use the framework's `Loader`, `Parser`, and `Splitter` components to ingest documents, then use an `Embedding` model and a `VectorStore` to create a searchable knowledge base.

Core Features of Agents-Flex

• Broad LLM Support: Natively supports a vast range of LLMs, including OpenAI, Google Gemini, Ollama, Deepseek, Alibaba Qwen, iFlytek Spark, Coze, and many others. It also allows easy integration of private models via HTTP/SSE/WS protocols.
• Powerful Function Calling: An intuitive, annotation-based system for defining and managing tools that LLMs can invoke, enabling the creation of task-oriented agents.
• Comprehensive RAG Components: Includes modules for document loading (from web, local files, databases), parsing various formats, and intelligent text splitting.
• Extensive Vector Store Integration: Seamlessly connects with numerous vector databases such as Milvus, Elasticsearch, Pgvector, Qdrant, Redis, OpenSearch, and more, providing flexibility for your RAG architecture.
• Flexible Memory Management: Features `MessageMemory` for maintaining conversational history and `ContextMemory` for tracking the state within complex execution chains.
• Sophisticated Agent & Chain Orchestration: Supports synchronous, asynchronous, and looping execution chains, allowing developers to build complex, multi-step workflows and autonomous agents (e.g., ReAct agents).
• Multi-Modal Capabilities: Extends beyond text to support interactions involving images and audio, enabling the development of richer AI applications.

Use Cases for Agents-Flex

Agents-Flex is suitable for a wide variety of AI application development scenarios:

• Intelligent Chatbots and Virtual Assistants: Build conversational agents that can remember past interactions and access external tools.
• Enterprise Knowledge Base Q&A: Create powerful RAG systems that allow users to ask questions in natural language and receive answers sourced from internal company documents.
• Automated Task Execution: Develop autonomous agents that can understand a user's goal, break it down into steps, and execute those steps by calling various APIs and local functions.
• Content Generation Tools: Integrate text, image, or audio generation capabilities directly into your Java applications.
• Data Processing and Analysis: Use LLMs to extract, summarize, and analyze information from large volumes of unstructured text.

Advantages of Agents-Flex

• Simplicity and Ease of Use: Its API is designed to be concise and intuitive, significantly reducing development time.
• Framework Agnostic: Unlike competitors like Spring-AI, it is not locked into a specific ecosystem. It works with any Java framework (Spring, Solon, etc.) or even in a plain Java application.
• Low System Requirements: Requires only JDK 8+, making it accessible for a wider range of projects, including legacy systems.
• Highly Extensible: Every core component is interface-based, encouraging customization and extension.
• Excellent Support for Chinese Models: Provides first-class support for many popular Chinese LLMs and services, a key advantage for developers in that market.
• Open Source: The framework is completely free, open-source, and benefits from active development and community contributions.

Pricing and Plans

Agents-Flex is an open-source project distributed under a permissive license. It is completely free to use for both personal and commercial projects. There are no paid plans or enterprise editions. The project is maintained on Gitee and GitHub, and contributions from the community are welcome.