SpringBoot Gemini API Consumer – Documentation

Purpose and Scope

This document provides an overview of the springboot-gemini-api-consumer repository, an application developed in Spring Boot that serves as an integration layer between client applications and the Google Gemini AI API.

The system exposes a REST interface that accepts prompts, transforms them into the format required by Gemini, manages authentication through API Keys, and returns the generated response.

• Main Goal: act as a secure and simple proxy to consume Gemini from frontend applications.
• Audience: frontend developers, backend developers, and architects who want to integrate Gemini into their systems.

For interactive testing, access the Swagger documentation: Swagger UI

---

System Description

The springboot-gemini-api-consumer is a Java microservice that:

• Receives prompt requests in JSON format.
• Builds the required hierarchy (Prompt → Contents → Parts).
• Sends the request to the Gemini endpoint (generativelanguage.googleapis.com).
• Returns the response to the client in JSON format.

The system is configured with CORS to accept requests from http://localhost:3000, enabling seamless integration with frontend applications.

---

High-Level System Architecture

Component Overview

• Client (Frontend / Console / Postman) → sends a JSON { "text": "..." }.
• Web Layer (ApiController) → receives the request and builds the Prompt DTO.
• Service Layer (AppService, GeminiService) → coordinates logic and makes the HTTP call to the Gemini service.
• Gemini API → processes the prompt and returns the response.
• ApiController → sends the response back to the client.

---

Technology Stack

Component	Technology	Version	Purpose
Runtime	Java	21	Execution environment
Framework	Spring Boot	3.4.5	Framework and DI
AI Integration	Spring AI BOM	1.0.0-M7	Abstractions for AI
Web Layer	Spring Boot Starter Web	3.4.5	REST API
Build Tool	Maven	3.x	Dependency management
Code Reduction	Lombok	Latest	Boilerplate reduction
Testing	Spring Boot Starter Test	3.4.5	Unit and integration testing

---

Core Components and Code Mapping

Application Layer

• SpringbootGeminiApiConsumerApplication

  • Entry point.
  • Exposes a RestTemplate bean for HTTP communication.

Web Layer

• ApiController

  • Exposes REST endpoints.

  • Main endpoint:

    POST /api/get-result-BreinLogic
    

  • CORS configured for http://localhost:3000.

  • Builds the Prompt → Contents → Parts DTO hierarchy.

Service Layer

• AppService

  • Coordinates business logic.
  • Receives Prompt, sends it to GeminiService, and returns the response.

• GeminiService

  • Calls Gemini endpoint directly.
  • Injects Gemini URL from application.properties.
  • Uses RestTemplate for POST requests.

---

Request Processing Flow

1. The client sends a JSON request:

   { "text": "Explain the Singleton pattern" }

2. ApiController receives the Parts, wraps it into a Contents, and then into a Prompt.

3. AppService coordinates the call and forwards it to GeminiService.

4. GeminiService makes a POST request to the Gemini endpoint.

5. The generated response is received and returned to the client.

---

Key Features

• API Proxy Pattern: acts as an intermediary between frontend and Gemini.
• Cross-Origin Support: CORS setup to allow requests from localhost:3000.
• API Key Management: Gemini key externalized via environment variables.
• Hierarchical DTO: Prompt → Contents → Parts model for complex prompts.
• Spring AI Integration: abstractions for AI integration.
• RESTful Interface: clean and simple API for integration.

---

Example Request & Response

Request

curl -X POST https://springboot-gemini-api-consumer.onrender.com/api/get-result-BreinLogic \
  -H "Content-Type: application/json" \
  -d '{"text": "Hi, what is your name?"}'

Response

{
  "response": "My name is BreinLogic. I am a friendly assistant and can answer questions about programming or general topics."
}

---

Future Improvements

• Dynamic personality configuration (e.g., BreinLogic as a friendly, general-purpose assistant).
• Multi-model support (Gemini 1.5 Flash, Gemini 1.5 Pro, etc.).
• Conversational context management (message history).
• Database integration for prompt logging and analysis.