LnYo-Cly/ai4j

Java 8+ agentic SDK: unified LLM access (OpenAI/Anthropic/DashScope/Doubao/DeepSeek...),unified AI service(image/video/audio...), Tool Calling, MCP, RAG, SKILLS,Agent Runtime, A2A, agent blueprint, agent harness, and a built-in Coding Agent CLI/TUI/ACP.

What it solves

ai4j is a comprehensive Java SDK designed to simplify the development of AI agents. It solves the problem of fragmented AI service integration by providing a unified interface for over 12 major model platforms (such as OpenAI, Anthropic, and DeepSeek) and offering built-in tools for agent orchestration, RAG, and long-term task management, which are typically missing or fragmented in the Java ecosystem.

How it works

The SDK provides a unified AiService factory that allows developers to switch between different LLM providers by simply changing a PlatformType configuration. It implements several agent execution modes including ReAct, CodeAct, and Deep Research, and uses a StateGraph for complex orchestration with conditional branching and loops.

For long-term tasks, it introduces ai4j-harness, a runtime that persists task states to storage (File/JDBC), allowing agents to be paused, resumed, and audited. It also includes a full RAG pipeline—from document loading (Tika/MinerU) and chunking to integration with eight different vector databases—all contained within the SDK.

Who it’s for

Java developers (JDK 8+) who want to build sophisticated AI agents, RAG applications, or long-running automated workflows that require persistence, auditability, and integration with multiple LLM providers within a Spring Boot or standalone Java environment.

Highlights

  • Unified Model Access: Single API for 12+ platforms including OpenAI, Anthropic, DeepSeek, and Ollama.
  • Agent Orchestration: Supports ReAct, CodeAct, and Deep Research modes with multi-agent team collaboration.
  • ai4j-harness: Transforms one-off AI calls into durable, recoverable, and auditable long-term tasks.
  • Built-in RAG: Complete pipeline including document parsing, hybrid search, and support for 8+ vector stores (e.g., Pinecone, Milvus, Qdrant).
  • Extensible Plugin System: Allows injection of custom tools, slash commands, and skills via Maven jars.
  • Ecosystem Connectivity: Supports Model Context Protocol (MCP) and A2A protocols for agent-to-agent collaboration.
  • Coding Agent: Includes a ready-to-use CLI/TUI/ACP tool for local coding assistance.

Related

  • Project
    agentscope-ai/agentscope-java

    A production-ready Java framework for building distributed, enterprise-grade AI agents with built-in support for session recovery, secure sandboxing, and multi-agent orchestration.

  • Project
    luoguoxiong/aipack

    aipack is a TypeScript‑first, lightweight AI‑agent framework. It includes its own runtime, a Tapable‑based plugin system, and built‑in adapters for 13+ LLM providers. Features include streaming vs. batch execution, persistent session storage, permission‑based tool approval, hybrid BM25 + vector long‑term memory, five‑stage context compression, multi‑agent orchestration (graph, pipeline, router, debate, map‑reduce), two‑way MCP integration, full observability (metrics, dashboard), and an offline evaluation suite. The project ships a CLI assistant, a library for embedding in apps, and example apps demonstrating real‑world use cases.

  • Project
    w8123/EnterpriseAgentFramework

    ReachAI is an open‑source Java‑centric platform that retrofits existing enterprise apps (OA, ERP, CRM, HR, etc.) with AI assistants. It scans UI and APIs, registers business methods as annotated Capabilities, uses LLM‑driven “AI Coding” to implement suggested refactors, and deploys Agents and versioned Workflows that can answer natural‑language queries, fill forms and trigger transactions—all while preserving original authentication, requiring user confirmation for writes, and providing full audit traces.

  • Project
    nageoffer/ragent

    Ragent AI is a production‑ready Java platform for building Agentic RAG applications. It provides hybrid retrieval (vector, keyword, graph, web), advanced query understanding, model tiering with fault‑tolerance, multi‑level session memory, Redis‑based traffic protection, and a full ingestion‑to‑answer loop. The code is split into seven Maven modules (infra‑ai, system, rag, agent, bootstrap, framework, mcp‑server) that isolate AI vendor, storage, and retrieval concerns, making components interchangeable. v2 uses a ReAct‑style agent loop with MCP tool calls, Skills, and human‑in‑the‑loop confirmations. Production features include circuit‑breaker, transactional RocketMQ, idempotency, tracing (Langfuse optional), SSE streaming, and Sa‑Token security. A React UI offers both chat with citation/agent view and an admin console for knowledge‑base and intent‑tree management. Targeted at Java backend engineers and interview‑preparation candidates who need a real‑world, extensible AI stack rather than a toy demo.

  • Project
    openai/openai-java

    Official Java client library for OpenAI’s REST API, offering sync/async, streaming, workload‑identity auth, and optional Bedrock support.