Spring AI

The Spring AI project provides a Spring-friendly API and abstractions for developing AI applications.

Its goal is to apply to the AI domain Spring ecosystem design principles such as portability and modular design and promote using POJOs as the building blocks of an application to the AI domain.

At its core, Spring AI addresses the fundamental challenge of AI integration: Connecting your enterprise Data and APIs with the AI Models.

The project draws inspiration from notable Python projects, such as LangChain and LlamaIndex, but Spring AI is not a direct port of those projects. The project was founded with the belief that the next wave of Generative AI applications will not be only for Python developers but will be ubiquitous across many programming languages.

You can check out the blog post Why Spring AI for additional motivations.

This is a high level feature overview. You can find more details in the Reference Documentation

• Support for all major AI Model providers such as Anthropic, OpenAI, Microsoft, Amazon, Google, and Ollama. Supported model types include:
  • Chat Completion
  • Embedding
  • Text to Image
  • Audio Transcription
  • Text to Speech
  • Moderation
• Portable API support across AI providers for both synchronous and streaming API options are supported. Access to model-specific features is also available.
• Structured Outputs - Mapping of AI Model output to POJOs.
• Support for all major Vector Database providers such as Apache Cassandra, Azure Vector Search, Chroma, Milvus, MongoDB Atlas, Neo4j, Oracle, PostgreSQL/PGVector, PineCone, Qdrant, Redis, and Weaviate.
• Portable API across Vector Store providers, including a novel SQL-like metadata filter API.
• Tools/Function Calling - permits the model to request the execution of client-side tools and functions, thereby accessing necessary real-time information as required.
• Observability - Provides insights into AI-related operations.
• Document injection ETL framework for Data Engineering.
• AI Model Evaluation - Utilities to help evaluate generated content and protect against hallucinated response.
• ChatClient API - Fluent API for communicating with AI Chat Models, idiomatically similar to the WebClient and RestClient APIs.
• Advisors API - Encapsulates recurring Generative AI patterns, transforms data sent to and from Language Models (LLMs), and provides portability across various models and use cases.
• Support for Chat Conversation Memory and Retrieval Augmented Generation (RAG).
• Spring Boot Auto Configuration and Starters for all AI Models and Vector Stores - use the start.spring.io to select the Model or Vector-store of choice.

Getting Started

Please refer to the Getting Started Guide for instruction on adding your dependencies.

Project Resources

• Documentation
• Issues

• Awesome Spring AI - A curated list of awesome resources, tools, tutorials, and projects for building generative AI applications using Spring AI

Breaking changes

• Refer to the upgrade notes to see how to upgrade to 1.0.0.M1 or higher.

Cloning the repo

This repository contains large model files. To clone it you have to either:

• Ignore the large files (won't affect the spring-ai behaviour) : GIT_LFS_SKIP_SMUDGE=1 git clone [email protected]:spring-projects/spring-ai.git.
• Or install the Git Large File Storage before cloning the repo.

Building

To build with running unit tests

./mvnw clean package

To build including integration tests.

./mvnw clean verify -Pintegration-tests

Note that you should set API key environment variables for OpenAI or other model providers before running. If the API key isn't set for a specific model provider, the integration test is skipped.

To run a specific integration test allowing for up to two attempts to succeed. This is useful when a hosted service is not reliable or times out.

./mvnw -pl vector-stores/spring-ai-pgvector-store -Pintegration-tests -Dfailsafe.rerunFailingTestsCount=2 -Dit.test=PgVectorStoreIT verify

Integration Tests

There are many integration tests ,so it often isn't realistic to run them all at once.

A quick pass through the most important pathways that runs integration tests for

• OpenAI models
• OpenAI autoconfiguration
• PGVector
• Chroma

can be done with the profile -Pci-fast-integration-tests and is used in the main CI build of this project.

A full integration test is done twice a day in the Spring AI Integration Test Repository

One way to run integration tests on part of the code is to first do a quick compile and install of the project

./mvnw clean install -DskipTests -Dmaven.javadoc.skip=true

Then run the integration test for a specifi module using the -pl option

./mvnw verify -Pintegration-tests     -pl spring-ai-spring-boot-autoconfigure  

Documentation

To build the docs

./mvnw -pl spring-ai-docs antora

The docs are then in the directory spring-ai-docs/target/antora/site/index.html

To reformat using the java-format plugin

./mvnw spring-javaformat:apply

To update the year on license headers using the license-maven-plugin

./mvnw license:update-file-header -Plicense

To check javadocs using the javadoc:javadoc

./mvnw javadoc:javadoc -Pjavadoc

To build with checkstyles enabled. Checkstyles are currently disabled, but you can enable them by doing the following:

./mvnw clean package -DskipTests -Ddisable.checks=false