Integrate OpenAI and Anthropic in Laravel 13: Build AI-Powered Applications with Multiple LLM Providers

Integrate OpenAI and Anthropic in Laravel 13

Artificial Intelligence is rapidly becoming a core feature of modern web applications. Instead of relying on a single AI provider, many production applications now integrate multiple Large Language Models (LLMs) to improve reliability, reduce costs, and choose the best model for each task.

Laravel 13 makes this process straightforward with its clean service architecture, HTTP client, dependency injection, configuration management, and environment variables.

In this guide, we'll build a reusable AI service capable of communicating with both OpenAI GPT and Anthropic Claude, allowing your application to switch providers with minimal code changes.

Why Integrate Multiple AI Providers?

Using multiple AI providers offers several advantages:

• High availability with automatic provider fallback
• Better cost optimisation
• Access to specialised models
• Reduced vendor lock-in
• Compare responses from different LLMs
• Future-proof AI architecture
• Better performance for different workloads

For example:

• GPT-5.5 for code generation
• Claude for long-document analysis
• Different providers for different pricing requirements

Prerequisites

Before starting, you'll need:

• Laravel 13
• PHP 8.3+
• Composer
• OpenAI API Key
• Anthropic API Key

Configuration

Create a configuration file.

// config/ai.php

return [

    'provider' => env('AI_PROVIDER', 'openai'),

    'openai' => [
        'key' => env('OPENAI_API_KEY'),
        'model' => 'gpt-5.5',
    ],

    'anthropic' => [
        'key' => env('ANTHROPIC_API_KEY'),
        'model' => 'claude-sonnet-4',
    ],

];

AI Service Example

Laravel's service layer keeps AI integrations clean and reusable.

namespace App\Services;

use Illuminate\Support\Facades\Http;

class AIService
{
    public function generate(string $prompt)
    {
        return config('ai.provider') === 'openai'
            ? $this->openai($prompt)
            : $this->anthropic($prompt);
    }

    protected function openai($prompt)
    {
        return Http::withToken(config('ai.openai.key'))
            ->post('https://api.openai.com/v1/responses', [
                'model' => config('ai.openai.model'),
                'input' => $prompt,
            ])
            ->json();
    }

    protected function anthropic($prompt)
    {
        return Http::withHeaders([
            'x-api-key' => config('ai.anthropic.key'),
            'anthropic-version' => '2023-06-01',
        ])
        ->post('https://api.anthropic.com/v1/messages', [
            'model' => config('ai.anthropic.model'),
            'max_tokens' => 1000,
            'messages' => [
                [
                    'role' => 'user',
                    'content' => $prompt,
                ]
            ]
        ])
        ->json();
    }
}

Using the Service

use App\Services\AIService;

public function generate(AIService $ai)
{
    return $ai->generate(
        "Explain Laravel Service Container"
    );
}

Benefits of This Architecture

Instead of scattering API calls throughout controllers, all AI communication is centralised.

Advantages include:

• Cleaner controllers
• Easier testing
• Provider switching without changing business logic
• Better scalability
• Centralised error handling
• Reusable across the entire application