PHPCrafters

AI in PHP: Integration Guide

AI in PHP: Integration Guide

The digital landscape is undergoing a seismic shift, driven by the pervasive integration of Artificial Intelligence (AI) and Machine Learning (ML). For the vast ecosystem of PHP, a language that powers over 75% of all websites, this evolution is not just an opportunity but a necessity to remain relevant and powerful in the modern web. The journey of infusing PHP applications with intelligence is a strategic architectural endeavor, offering multiple pathways, each with distinct trade-offs in control, complexity, cost, and performance. This guide provides a deep dive into the methods, tools, and best practices for seamlessly integrating AI and ML into PHP, empowering developers to build the next generation of intelligent web applications.

Architectural Approaches to AI Integration in PHP

The foundational decision in any AI-integrated PHP project is the choice of architecture. This choice fundamentally dictates the application’s capabilities, operational overhead, and long-term maintainability. The spectrum of architectural patterns ranges from leveraging powerful external services to building self-sufficient systems within the native PHP environment.

The External API Model: Democratizing AI with Cloud Power

The most pragmatic and widely adopted approach for PHP applications, especially those built on modern frameworks like Laravel and Symfony, is the integration of external AI APIs. This model allows developers to tap into the immense computational resources and state-of-the-art models developed by specialized AI companies like OpenAI, Google, and Anthropic.

By making simple HTTP requests to well-defined endpoints, a PHP application can access a staggering array of advanced functionalities, including Natural Language Processing (NLP), generative text, image recognition, and audio transcription. This “external service” model has become the de facto standard for rapidly implementing features like AI-powered chatbots, content summarization, sentiment analysis, and automated code generation.

Key Advantages:

  • Speed and Ease of Implementation: Developers can prototype and deploy sophisticated AI features in days or even hours by consuming existing APIs, abstracting away the immense complexity of ML infrastructure, data pipelines, and model training.
  • Access to Cutting-Edge Models: API providers continuously update their models, giving applications immediate access to the latest advancements in AI without any development overhead.
  • Scalability: The cloud infrastructure of providers like OpenAI automatically scales to handle demand, freeing developers from capacity planning.

Inherent Challenges:

  • Third-Party Dependency: The application’s functionality becomes dependent on the reliability and availability of an external service.
  • Cost Management: Usage is typically metered (e.g., per token), requiring careful monitoring to avoid unexpected costs.
  • Network Latency: Every inference call involves a round-trip over the network, which can impact response times.
  • Data Privacy: Sensitive data must be sent to a third-party server, which may be a concern under strict data sovereignty regulations.

The Native Library Model: The Path to Self-Sufficiency

In direct contrast to the external API model, a growing and maturing ecosystem of native PHP machine learning libraries offers a path toward complete control and independence. Libraries like Rubix ML, PHP-ML, and the revolutionary Transformers PHP enable developers to build, train, and deploy ML models entirely within the PHP environment.

This approach is particularly compelling for scenarios requiring offline functionality, ultra-low-latency inference, or where strict data privacy mandates prohibit data from leaving the organization’s infrastructure.

Key Advantages:

  • Unparalleled Control: Developers have full control over the entire ML pipeline, from data preprocessing and feature engineering to model deployment and inference.
  • Data Privacy and Sovereignty: All data remains within the application’s own infrastructure, a critical requirement for healthcare, finance, and other regulated industries.
  • Offline Operation: Applications function completely independently of an internet connection.
  • Performance: For inference tasks, eliminating network latency can lead to significantly faster response times, especially for real-time applications.

Inherent Challenges:

  • Developer Complexity: The team is responsible for the full ML lifecycle, which necessitates a deeper understanding of machine learning principles.
  • Computational Limits: While inference is efficient, the computationally intensive task of training complex models from scratch is generally better suited to Python-based environments.
  • Ecosystem Maturity: While rapidly improving, the PHP ML ecosystem does not yet match the depth and breadth of Python’s, particularly in areas like widespread GPU acceleration.

The Hybrid Model: Balancing Power and Control

In sophisticated, real-world applications, a third, hybrid approach often emerges as the optimal strategy. This model thoughtfully combines the strengths of both external and internal systems.

A PHP application might use an external API like OpenAI for computationally expensive or highly specialized generative tasks (e.g., summarizing a lengthy legal document or creating marketing images) while relying on a native PHP library like Rubix ML for real-time, low-latency predictive tasks (e.g., classifying user feedback sentiment or detecting fraudulent transactions in real-time).

This hybrid strategy allows architects to leverage the best tool for each specific job, balancing the raw power and sophistication of external models with the efficiency, control, and privacy of native execution. It requires careful design to manage data flow, synchronous vs. asynchronous calls, and consistent error handling across disparate systems.

Deep Dive into External API Integration: The OpenAI Ecosystem

For developers seeking to rapidly infuse PHP applications with advanced AI, the OpenAI API stands as a leading, extensively documented solution. Its integration is a well-established pattern that leverages Large Language Models (LLMs) like GPT-4 and GPT-3.5-turbo.

Authentication and Setup

The process begins by obtaining an API key from the OpenAI dashboard. This key is a sensitive secret and must be stored securely using environment variables (e.g., OPENAI_API_KEY), loaded into the PHP application at runtime using a library like vlucas/phpdotenv. This practice ensures credentials are separate from application code, enhancing security across different deployment environments.

Making API Requests

Communication with the OpenAI API is conducted through HTTP POST requests. For chat functionality, the relevant endpoint is https://api.openai.com/v1/chat/completions. Developers commonly use the Guzzle HTTP client, installed via Composer, to make these requests.

A request must include:

  • Headers: Content-Type: application/json and Authorization: Bearer <YOUR_API_KEY>.
  • JSON Body: The core of the request, specifying the model (e.g., gpt-3.5-turbo), a message array, and parameters like max_tokens and temperature.

The message array uses role-based pairs (system, user, assistant) to structure the conversation, allowing the application to maintain a coherent, multi-turn dialogue.

Beyond Chat: A Versatile Toolkit

The OpenAI API’s versatility extends far beyond simple chat. PHP developers can leverage it for:

  • Content Generation: Crafting jokes, poems, and summaries.
  • Image Generation: Creating images from text descriptions using DALL-E.
  • Audio Processing: Transcribing speech to text with the Whisper model.
  • Embeddings: Generating numerical representations of text for semantic search and recommendation engines.

Framework Integration and Best Practices

Within the Laravel ecosystem, packages like laravel-openai provide a clean, facade-based interface, abstracting away HTTP client logic. Other packages like prism.php offer a unified interface for multiple AI providers, reducing vendor lock-in.

To manage the challenges of external APIs, developers should adopt key strategies:

  • Cost Control: Use the tokenizer tool to estimate costs and monitor usage closely.
  • Latency Reduction: Reduce max_tokens, use smaller models for simpler tasks, and implement caching (e.g., with Redis) to store common API responses.
  • Reliability: Implement retry logic with exponential backoff, set appropriate timeouts, and have fallback mechanisms for when the API is unavailable.
  • Output Handling: Be prepared for model “hallucinations” and invalid JSON by using prompt engineering and carefully parsing API responses.

Leveraging Native PHP Machine Learning Libraries

For developers prioritizing control and data privacy, the native PHP ML ecosystem offers a powerful and increasingly capable toolkit.

Rubix ML: A High-Level, Production-Ready Library

Rubix ML is a comprehensive, free open-source library supporting the entire machine learning lifecycle. It features over 40 supervised and unsupervised learning algorithms for classification, regression, clustering, and anomaly detection. Its clean, intuitive API encourages the use of pipelines to streamline workflows from data loading to model evaluation. A key feature is model persistence, allowing trained models to be saved and reloaded for inference, making it production-ready. It requires PHP 7.4+ and can be installed via Composer.

PHP-ML: General-Purpose Machine Learning

PHP-ML is an established library focused on providing a suite of standard ML algorithms. It supports classification methods like K-Nearest Neighbors and Support Vector Machines, regression, clustering, and neural networks. It includes tools for cross-validation and preprocessing, making it a versatile choice for traditional ML tasks. A demonstrated use case involves classifying email attachment images by extracting custom features, highlighting its utility in computer vision.

Transformers PHP: Bringing Modern AI to PHP

Transformers PHP is a revolutionary library that enables the execution of complex transformer models (like BERT and GPT-2) directly within PHP. It leverages the ONNX Runtime via PHP’s Foreign Function Interface (FFI), allowing it to achieve near-native performance for deep learning inference. The library downloads and caches models from the Hugging Face Hub locally, enabling fully offline operation. It supports a vast array of tasks, including sentiment analysis, translation, and text generation, effectively bringing the power of the Hugging Face ecosystem to PHP.

The Broader Ecosystem and Considerations

The landscape also includes specialized tools like PHPTensorFlow (TensorFlow bindings), Fann (neural networks), and phpInsight (sentiment analysis). The overall maturity of the native PHP ML ecosystem is advancing rapidly. However, developers should be aware that training is best done in Python, with PHP excelling at inference. Furthermore, while performance is impressive with FFI and ONNX, widespread GPU support remains a limitation compared to Python.

Frameworks as Enablers: AI Integration in Laravel and Symfony

The choice of PHP framework profoundly shapes the AI integration process, influencing architectural patterns, dependency management, and deployment.

Laravel: Rapid Development and a Rich Ecosystem

Laravel’s “batteries-included” philosophy and gentle learning curve make it ideal for rapidly prototyping and integrating AI features, particularly those using external APIs. Its ecosystem is rich with community packages like laravel-openai and prism.php that drastically simplify integration.

Laravel’s architectural patterns are inherently conducive to AI:

  • Queues: The robust queue system is perfect for offloading long-running AI tasks (e.g., report generation) to background workers, keeping the application responsive.
  • Service Container: AI logic can be encapsulated within dedicated service classes, promoting separation of concerns, testability, and maintainability.

Symfony: Enterprise-Grade Scalability and Control

Symfony is a meta-framework renowned for its modularity, scalability, and enterprise-grade architecture. Its component-based design offers granular control, making it an excellent choice for large-scale, mission-critical AI systems.

A landmark development is the official symfony/ai-bundle, which provides a standardized way to integrate with AI platforms, manage vector stores for Retrieval-Augmented Generation (RAG), and orchestrate AI agents. Symfony’s powerful, declarative configuration system (using YAML and environment variables) offers a highly organized and secure way to manage API keys and service configurations across environments.

Choosing the Right Framework

The decision is strategic:

  • Choose Laravel for projects demanding rapid development, quick iteration, and seamless integration with third-party AI services. It is ideal for startups and MVPs.
  • Choose Symfony for large-scale, complex applications requiring architectural rigor, deep customization, and long-term maintainability. It is the preferred choice for enterprise-grade systems with complex AI workloads.

Advanced Patterns for Interactive and Production-Ready AI

Building professional, production-ready AI applications requires moving beyond basic API calls to adopt advanced patterns for user experience and resilience.

Real-Time Response Streaming

For interactive applications like chatbots, waiting for a complete LLM response creates a poor user experience. Real-time streaming delivers the response progressively, creating a “typing effect.” The OpenAI API supports this via Server-Sent Events (SSE).

In a Laravel application, this can be managed using the response()->stream() helper. The server-side script processes tokens from OpenAI as they are generated and sends them as SSE events to the client. JavaScript on the client side, using the EventSource API, listens for these events and appends each chunk to the display. This requires configuring headers to prevent buffering and results in a significantly more engaging user experience.

Security, Reliability, and Testing

Security:

  • API Keys: Always store keys in environment variables, never in code.
  • OAuth 2.0: For user-level authentication with third-party services, use OAuth instead of handling credentials.
  • Input Sanitization: Validate and sanitize all user prompts to prevent prompt injection attacks that could manipulate the model.

Reliability:

  • Defensive Coding: Wrap API calls in try-catch blocks to handle network errors and rate limits gracefully.
  • Retry Logic: Implement retry mechanisms with exponential backoff for transient failures.
  • Rate Limiting: Enforce application-level rate limits to prevent abuse and stay within provider quotas.
  • Output Validation: Assume model outputs can be erroneous. Use prompt engineering to request structured data (like JSON) and implement server-side validation logic.

Testing:
Testing AI features is challenging due to their non-deterministic nature. The solution is to mock the AI API. In Laravel, the Http::fake() method can simulate API responses, returning predefined JSON fixtures. This allows for deterministic unit tests that verify application logic without incurring cost or latency.

Strategic Considerations and Future Outlook

Integrating AI into PHP is a strategic decision with long-term implications. The ecosystem, while powerful, has considerations.

Current Landscape:

  • Division of Labor: The most effective strategy often involves using Python for training complex models and PHP (with Rubix ML or Transformers PHP) for efficient, low-latency inference.
  • GPU Acceleration: Broad, standardized GPU support for training and accelerating inference in PHP remains a bottleneck, though CPU performance with FFI and ONNX is highly capable.

Future Trends:

  1. Local and Private Model Execution: Packages like laravel-ollama signify a strong shift towards running open-source LLMs locally, addressing data privacy, security, and cost concerns for regulated industries.
  2. AI Agents and RAG: Frameworks are beginning to provide abstractions for building autonomous AI agents that can reason and use tools. Retrieval-Augmented Generation (RAG), which enhances LLMs with information from a knowledge base, is becoming a standard workflow, facilitated by integration with vector databases.

Conclusion

The integration of AI and ML into PHP applications is a multifaceted and strategic endeavor that unlocks immense potential for innovation. The path forward is not a binary choice but a pragmatic blend: leveraging powerful external APIs for complex generative tasks while utilizing the growing capabilities of native PHP libraries for efficient, private, and low-latency inference.

The choice between Laravel and Symfony will set the architectural foundation, guiding the development velocity and scalability of the project. By embracing advanced patterns like real-time streaming and adhering to rigorous best practices in security, reliability, and testing, PHP developers can transition from simple integrations to building robust, professional, and intelligent applications. The continued evolution of the PHP AI ecosystem promises to further democratize this powerful technology, cementing PHP’s role as a vital and intelligent force in the future of the web.

References

  1. PHP-ML – Read the Docs. https://php-ml.readthedocs.io/
  2. Integrating Machine Learning Models into PHP Applications. https://medium.com/codex/integrating-machine-learning-models-into-php-applications-a-step-by-step-guide-a4a44cdde93f
  3. Empowering PHP Developers: The Definitive Guide to … https://www.rockersinfo.com/empowering-php-developers-the-definitive-guide-to-the-top-20-ai-libraries-for-seamless-integration/
  4. Machine Learning with PHP. https://dev.to/robertobutti/machine-learning-with-php-5gb
  5. Rubix ML: Welcome. https://rubixml.com/
  6. Build a News Classifier Using Rubix ML. https://dev.to/arafatweb/machine-learning-in-php-build-a-news-classifier-using-rubix-ml-e45
  7. Basic Introduction – Rubix ML. https://rubixml.github.io/ML/2.0/basic-introduction.html
  8. rubix-ml – GitHub Topics. https://github.com/topics/rubix-ml
  9. Rubix ML download. https://sourceforge.net/projects/rubix-ml.mirror/
  10. Introducing Rubix ML: PHP’s native machine learning library. https://www.linkedin.com/posts/vivekpaswan624_rubixml-machinelearning-php-activity-7356212781802487810-JiNU
  11. ideaguy3d/php-machine-learning: PHP-ML. https://github.com/ideaguy3d/php-machine-learning
  12. 12 Must-Know Libraries to Supercharge Your PHP … https://medium.com/@CodeMoox/12-must-know-libraries-to-supercharge-your-php-development-in-2025-7ac1f83a704a
  13. Machine Learning Using PHP-ML, implementing image … https://github.com/deltastateonline/ml.php
  14. How to Use OpenAI’s GPT-3.5-Turbo API with PHP. https://medium.com/@vishwaspinnawala/how-to-use-openais-gpt-3-5-turbo-api-with-php-19bffe4c98e7
  15. Working with the OpenAI API with Laravel and PHP. https://www.cowshedworks.co.uk/blog/working-with-the-openai-api-with-laravel-and-php
  16. Fun With OpenAI and Laravel, Ep 01 – Hello, AI. https://www.youtube.com/watch?v=eWRBCEpOlq8
  17. Integrating AI in PHP Apps: OpenAI, Chatbots, and Beyond. https://200oksolutions.com/blog/integrating-ai-in-php-apps-openai-chatbots-and-beyond/
  18. How to use OpenAI API on PHP? https://tech.jotform.com/how-to-use-openai-api-on-php-773b61d5ef83
  19. OpenAI Realtime API: A Guide With Examples. https://www.datacamp.com/tutorial/realtime-api-openai
  20. Transform Your PHP Projects with Powerful AI Data Insights. https://ai.plainenglish.io/transform-your-php-projects-with-powerful-ai-data-insights-99dc40ef2ed8
  21. How to use AI in PHP: Practical Applications and Examples. https://www.linkedin.com/posts/vinectpal468.php-ai-machinelearning-activity-7372976379161792513-TVe7
  22. orhanerday/open-ai: OpenAI PHP SDK. https://github.com/orhanerday/open-ai
  23. API Reference – OpenAI API. https://platform.openai.com/docs/api-reference/introduction
  24. GPT Action authentication – OpenAI API. https://platform.openai.com/docs/actions/authentication
  25. Authentication Methods | Security | Openai Api Tutorial. https://www.swiftorial.com/tutorials/artificial_intelligence/openai_api/security/authentication_methods
  26. API Security oAuth, BasicAuth, apikeys using PHP. https://www.youtube.com/watch?v=iEQ-3azEtuo
  27. How to Create and Use an OpenAI ChatGPT API Key. https://www.codecademy.com/article/creating-an-openai-api-key
  28. OAuth vs API Keys: Which API authentication method to … https://www.digitalapi.ai/blogs/oauth-vs-api-keys-which-api-authentication-method-to-choose
  29. Azure OpenAI in Microsoft Foundry Models REST API … https://learn.microsoft.com/en-us/azure/ai-foundry/openai/reference?view=foundry-classic
  30. The API Key security scheme in OpenAPI – Speakeasy. https://www.speakeasy.com/openapi/security/security-schemes/security-api-key
  31. Integrating OpenAI API with PHP: A Comprehensive Guide. https://iocoding.com/blog/integrating-openai-api-with.php/
  32. Production best practices – OpenAI API. https://platform.openai.com/docs/guides/production-best-practices
  33. Libraries – OpenAI API. https://platform.openai.com/docs/libraries
  34. openai.php/client. https://github.com/openai.php/client
  35. A PHP OpenAI API Client (SDK). https://community.openai.com/t/a-php-openai-api-client-sdk/22481
  36. An OpenAI API Client (SDK) for Your PHP Projects. https://tectalic.com/blog/openai-php-client-library-sdk/
  37. OpenAI PHP Client download. https://sourceforge.net/projects/openai-php-client.mirror/
  38. OpenAI PHP extension library – PHP 8.3. https://openai-php.com/
  39. OpenAI Libraries. https://notes.kodekloud.com/docs/Introduction-to-OpenAI/ Pre-Requisites/OpenAI-Libraries
  40. Overview – OpenAI API. https://platform.openai.com/
  41. Add AI features to your PHP apps – Agiledrop. https://www.agiledrop.com/blog/ add-ai-features-to-your-php-apps
  42. 8 Best AI for PHP: Top Tools for Faster Development. https://zencoder.ai/blog/best-ai-for-php
  43. Symfony vs CodeIgniter: Choosing the Best PHP Framework … https:// prevaj.com/symfony-vs-codeigniter-which-framework-is-best-for-your-project/
  44. Machine learning inference with PHP explained. https://upsun.com/blog/ml-inference-in-php-by-example/
  45. Symfony vs Laravel in 2025: Which PHP Framework Suits AI … https:// prevaj.com/symfony-vs-laravel-which-php-framework-should-you-use-in-the-age-of-ai/
  46. The Future of Laravel: AI & Machine Learning Integration in … https:// acquaintsoft.com/blog/ai-and-ml-integration-with-laravel
  47. Laravel vs Symfony: Which Framework Fits Your Project In … https:// webandcrafts.com/blog/laravel-vs-symfony
  48. Laravel in 2025: Your Go-To Guide to This PHP … https://medium.com/ @Jason4474/laravel-in-2025-your-go-to-guide-to-this-php-development-dynamo-346652d6c0b3
  49. Future of Laravel with AI: Building Smarter, Scalable Web … https:// www.cmarix.com/blog/future-of-laravel-with-ai/
  50. Top 9 Laravel AI Tools Every Developer Should Know in … https:// laracopilot.com/blog/laravel-ai-tools-for-developers-2025/
  51. Integrating Artificial Intelligence and Machine Learning with … https:// www.linkedin.com/pulse/integrating-artificial-intelligence-machine-learning-laravel-asith-8x2rc
  52. Laravel vs Symfony: The Best PHP Framework in 2025? https:// www.icoderzsolutions.com/blog/laravel-vs-symfony/
  53. How to Get Started With PHP AI in 2025 and Beyond? https:// www.linkedin.com/pulse/how-get-started-php-ai-2025-beyond-binstellar-technologies-pvt-ltd-awxff
  54. Why PHP is a Smart Pick for Web Development in 2025? https://medium.com/ @Jason4474/why-php-is-a-smart-pick-for-web-development-in-2025-e967af0ec0c5
  55. Harnessing AI in PHP Web Development: The 2025 Trend … https://www.maventricks.com/blog/harnessing-ai-in.php-web-development-the-2025-trend-to-watch/
  56. Streaming API responses. https://platform.openai.com/docs/guides/streaming-responses
  57. Realtime API. https://platform.openai.com/docs/guides/realtime
  58. ChatGPT Text Streaming: “Typing Effect” with OpenAI’s API … https://dekraker.medium.com/chatgpt-text-streaming-typing-effect-with-openais-api-and-javascript-php-ajax-85bcef4e60eb
  59. Seeking Simple PHP Example with ChatGPT ‘Stream’ … https://community.openai.com/t/seeking-simple.php-example-with-chatgpt-stream-functionality/277165
  60. OpenAI Realtime API: The Missing Manual. https://www.latent.space/p/realtime-api
  61. javascript – How do I Stream OpenAI’s completion API? https://stackoverflow.com/questions/73547502/how-do-i-stream-openais-completion-api
  62. Streaming OpenAI Responses in Laravel with Server-Sent … https://ahmadrosid.com/blog/laravel-openai-streaming-response
  63. RubixML/ML: A high-level machine learning and … https://github.com/RubixML/ML
  64. Rubix ML. https://github.com/RubixML
  65. RubixML/Sentiment. https://github.com/RubixML/Sentiment
  66. RubixML/Iris: The original lightweight introduction … https://github.com/RubixML/Iris
  67. The Best AI Packages for Laravel (2025 Edition). https://medium.com/@khouloud.haddad/the-best-ai-packages-for-laravel-2025-edition-e8e6da368e04
  68. Laravel AI Integration Tutorial: Complete Guide 2025. https://jetthoughts.com/blog/laravel-ai-integration-tutorial-complete-guide/
  69. Building a Laravel Package for Ollama – The full guide. https://dev.to/cammanhhoang/building-a-laravel-package-for-ollama-the-full-guide-laem
  70. cloudstudio/ollama-laravel. https://github.com/cloudstudio/ollama-laravel
  71. Use Ollama LLM Models Locally with Laravel. https://laravel-news.archivoh.com/archive/8-use-ollama-llm-models-locally-with-laravel
  72. Laravel PHP Requirements. https://www.zend.com/blog/laravel-php-requirements
  73. Using Docker to Solve PHP Version Compatibility Issues. https://medium.com/@hafizzeeshan619/using-docker-to-solve.php-version-compatibility-issues-a-practical-guide-6f2e9680dd99
  74. PHP: Supported Versions. https://www.php.net/supported-versions.php
  75. Key Differences Between PHP Versions 8.1, 8.2, 8.3, and 8.4. https://dhruvilblog.hashnode.dev/discover-the-key-differences-between-php-versions-81-82-83-and-84
  76. Machine Learning Inference in PHP by example. https://live.symfony.com/2024-vienna-con/schedule/machine-learning-inference-in.php-by-example-leverage-onnx-and-transformers-on-symfony
  77. ML inference in PHP by example leverage ONNX and … https://www.youtube.com/watch?v=KVEy29n2Ges
  78. AI Bundle (Symfony Docs). https://symfony.com/doc/current/ai/bundles/ai-bundle.html
  79. CodeWithKyrian/transformers.php. https://github.com/CodeWithKyrian/transformers.php
  80. Deep dive into ONNX and Integration with ML/DL … https://medium.com/@ravindusenaratne/deep-dive-into-onnx-and-integration-with-ml-dl-frameworks-96cd2d790c8a
  81. SymfonyCon Vienna 2024: Machine Learning Inference in … https://app.daily.dev/posts/symfonycon-vienna-2024-machine-learning-inference-in.php-by-example-leverage-onnx-and-transformers-fcrqj88tt
  82. Kicking off the Symfony AI Initiative. https://symfony.com/blog/kicking-off-the-symfony-ai-initiative
  83. Configuring Symfony (Symfony Docs). https://symfony.com/doc/current/configuration.html
  84. Exploring Dependency Injection in Symfony. https://codesignal.com/learn/courses/php-symfony-basics/lessons/exploring-dependency-injection-in-symfony