Migration Guide: Spring AI Google GenAI Autoconfiguration

Overview

This guide helps you migrate from the old Vertex AI-based autoconfiguration to the new Google GenAI SDK-based autoconfiguration.

Starter Dependencies

Spring AI provides separate starters for Google GenAI functionality:

Chat Functionality

<dependency>
    <groupId>org.springframework.ai</groupId>
    <artifactId>spring-ai-starter-model-google-genai</artifactId>
    <version>1.1.0-SNAPSHOT</version>
</dependency>

Embedding Functionality

<dependency>
    <groupId>org.springframework.ai</groupId>
    <artifactId>spring-ai-starter-model-google-genai-embedding</artifactId>
    <version>1.1.0-SNAPSHOT</version>
</dependency>

Note: If you need both chat and embedding capabilities, include both starters in your project. The starters are designed to be used independently or together based on your requirements.

Key Changes

1. Property Namespace Changes

Old properties:

spring.ai.vertex.ai.gemini.project-id=my-project
spring.ai.vertex.ai.gemini.location=us-central1
spring.ai.vertex.ai.gemini.chat.options.model=gemini-pro
spring.ai.vertex.ai.embedding.text.options.model=textembedding-gecko

New properties:

# For Vertex AI mode
spring.ai.google.genai.project-id=my-project
spring.ai.google.genai.location=us-central1
spring.ai.google.genai.chat.options.model=gemini-2.0-flash

# For Gemini Developer API mode (new!)
spring.ai.google.genai.api-key=your-api-key
spring.ai.google.genai.chat.options.model=gemini-2.0-flash

# Embedding properties
spring.ai.google.genai.embedding.project-id=my-project
spring.ai.google.genai.embedding.location=us-central1
spring.ai.google.genai.embedding.text.options.model=text-embedding-004

2. New Authentication Options

The new SDK supports both:

• Vertex AI mode: Using Google Cloud credentials (same as before)
• Gemini Developer API mode: Using API keys (new!)

3. Removed Features

• transport property is no longer needed
• Multimodal embedding autoconfiguration has been removed (pending support in new SDK)

4. Bean Name Changes

If you were autowiring beans by name:

• vertexAi → googleGenAiClient
• vertexAiGeminiChat → googleGenAiChatModel
• textEmbedding → googleGenAiTextEmbedding

5. Class Changes

If you were importing classes directly:

• com.google.cloud.vertexai.VertexAI → com.google.genai.Client
• org.springframework.ai.vertexai.gemini.* → org.springframework.ai.google.genai.*

Migration Steps

1. Update your application properties:

   • Replace spring.ai.vertex.ai.* with spring.ai.google.genai.*
   • Remove any transport configuration

2. If using API key authentication:

   • Set spring.ai.google.genai.api-key property
   • Remove project-id and location for chat (not needed with API key)

3. Update any custom configurations or bean references

4. Test your application thoroughly

Environment Variables

export GOOGLE_CLOUD_PROJECT=my-project
export GOOGLE_CLOUD_LOCATION=us-central1

New (additional option):

export GOOGLE_API_KEY=your-api-key

Backward Compatibility

The old autoconfiguration module is still available but deprecated. We recommend migrating to the new module as soon as possible.