Spring AI Alibaba Examples: runnable samples for building AI apps in Java

What the examples repository offers

Java developers adopting Spring AI, and its Alibaba-flavored extension Spring AI Alibaba, need worked code to see how the pieces fit. Spring AI Alibaba Examples is that: a collection of runnable samples demonstrating how to use Spring AI and Spring AI Alibaba across many scenarios. The repository is organized as example modules, a hello-world, chat and chat-memory, RAG, MCP, multi-model and multi-platform, agents, a graph example, NL2SQL, audio, image, evaluation, and an agent-to-agent example, so each common building block of an AI application has a self-contained sample. The audience is Java and Spring developers evaluating or learning this stack who want to run real examples rather than read reference docs alone. It is explicitly an examples repository, so its purpose is to teach the framework by demonstration, and you take patterns from it into your own application rather than importing it as a dependency.

A module per AI scenario

The structure is the value. Each numbered or named module targets one capability, so you can go straight to the sample for what you need: chat and chat-memory show conversation and history, the RAG example shows retrieval-augmented generation, the MCP example shows Model Context Protocol integration, the graph example shows composing multi-step flows, the NL2SQL example shows natural-language querying of a database, and audio and image examples show multimodal use, with an evaluation module for measuring outputs and a Python agent-to-agent example for cross-language agent communication. This breadth maps the surface of Spring AI Alibaba, so the repository doubles as a menu of what the framework can do. Because the samples are standard Maven Spring Boot projects, each runs on its own, and a docker-compose directory helps stand up supporting services such as vector stores that some examples require.

Running an example

Because this is a samples repository rather than a package, you clone it and run the module you want. Each example is a Maven-based Spring Boot application, so you build and run it with your usual Maven or IDE workflow, providing the model credentials that sample needs, typically Alibaba's model service or another provider Spring AI supports. Some examples depend on supporting infrastructure, a vector store for RAG, a database for NL2SQL, which the repository's docker-compose setup helps start, and the README notes that RAG, MCP and vector examples generally need those services running first. The first real use is running the hello-world or a basic chat example to confirm your credentials and environment work, then moving to the module matching your goal, RAG, MCP or NL2SQL, and starting its required services before running it.

Where a samples repository stops

The limitations are those of a teaching resource. These are examples meant to be read and adapted, not production code to deploy as-is, and they depend on the Spring AI and Spring AI Alibaba frameworks and on model providers, often Alibaba's, plus supporting services for some modules, so running everything requires credentials, infrastructure and cost. As a companion to fast-moving frameworks, individual samples can lag the current framework versions and may need adjustment to run against the latest releases, so pin the versions you learn against and check the framework docs when a sample misbehaves. And the documentation leans toward the Alibaba ecosystem, which shapes the default providers. None of this is a fault; it is what an examples repository is, and treating it as a curriculum and a pattern source rather than a dependency is the right approach.

The examples versus framework docs or another stack's samples

The alternatives are the Spring AI reference documentation, or samples for a different Java AI stack such as LangChain4j. Reference docs are authoritative and current but describe APIs rather than giving a runnable, end-to-end path, so you assemble the working example yourself. LangChain4j and its samples target a different framework, so they help only if you choose that stack. Spring AI Alibaba Examples' difference is that it is the official, runnable sample set for this specific framework and its Alibaba extension, covering many scenarios in one place, so a Spring developer sees exactly how chat, RAG, MCP, agents and the rest are built here. Use the reference docs to look up specifics and this repository to learn by running; for teams committed to Spring AI Alibaba, the official examples are the most direct way to become productive with it.

Apache-2.0 and maintenance

The repository is Apache-2.0 licensed, a permissive license, so you can freely reuse and adapt the sample code, which is appropriate for a learning resource. The last push was on 2026-07-01, and the breadth of modules, from hello-world through agents, graphs, NL2SQL and evaluation, shows the maintainers keeping pace with the framework's expanding capabilities. Because it accompanies actively developed frameworks, treat the samples as a snapshot: pin the Spring AI and Spring AI Alibaba versions you learn against, and if an example does not run, check it against the current framework documentation before assuming the sample is broken. Start with the hello-world or basic chat module to confirm your setup and credentials, then follow the module that matches your goal, standing up any required services from the docker-compose setup first.

Editorial conclusion

Work through Spring AI Alibaba Examples if you are a Java and Spring developer learning or evaluating Spring AI and Spring AI Alibaba and want runnable samples across chat, RAG, MCP, agents, NL2SQL and multimodal scenarios. Do not treat the examples as production code or expect them to run without model credentials and, for some modules, supporting services. Clone the repository, run the hello-world or basic chat module first to confirm your setup, then follow the module for your goal, starting its required services from docker-compose.

Frequently asked questions

Official sources

1. Issues
2. License: Apache-2.0
3. Project website
4. README
5. spring-ai-alibaba/examples on GitHub