langchain-java

Why we included this project

Java teams building LLM applications have usually had to reach for Python, since the ecosystem's tooling grew up around it. This project brings the familiar LangChain pieces, prompt templates, chains, chat models, and agents, into idiomatic Java, so a Spring or Spark shop can stay in its existing stack. Its real angle is the Big Data side: it ships toolkits that let an agent query Spark SQL and Flink SQL, which matters for data-platform teams who want natural-language access to their warehouse without standing up a separate Python service. The library covers the expected basics, OpenAI and Azure OpenAI clients, Ollama and ChatGLM for local models, and Pinecone and Milvus vector stores for retrieval. If you are weighing a JVM-native alternative to Python LangChain for RAG, SQL agents, or summarization, the worked examples in the repo give a concrete feel for the API before you commit.