11 Best Laravel AI Tools and Packages to Use in 2026

The best Laravel AI tools in 2026 include the official Laravel AI SDK (laravel/ai) for adding AI features to your app, Laravel Boost for making coding agents Laravel-aware, Prism for multi-provider LLM calls, and StackMivo for generating complete Laravel apps from a prompt. The rest of this list covers specialist packages for OpenAI, Gemini, local models, translation, token counting, and machine learning.

The Laravel AI landscape has changed significantly over the past year. In September 2026, the Laravel team released version 1.0 of its own AI SDK, so you no longer need a third-party package just to call an LLM. That's great news, but it also makes choosing the right tool harder: which tools overlap, and which ones still deserve a place in your stack?

This 2026 update compares 11 Laravel AI tools side by side, explains the new laravel/ai package in detail, and gives you a simple way to choose the right tool for each job.

Key Takeaways

• laravel/ai is now the default choice for AI features inside Laravel apps. It's official, free under the MIT license, reached v1.0 on September 23, 2026, and requires PHP 8.3+ with Laravel 12 or 13.

• Laravel AI tools fall into three groups: runtime SDKs for AI features your users interact with, developer tools that help you write code, and app builders that generate entire applications.

• Prism and openai-php/laravel are still solid choices when you need provider-specific features or already use them in production.

• Laravel Boost and StackMivo solve different problems: Boost makes your coding agent smarter, while StackMivo generates a full Laravel codebase.

• For machine learning in pure PHP, use Rubix ML. The older PHP-ML library is no longer maintained.

The developers in this article's examples are illustrative composites based on common Laravel scenarios.

Laravel AI Tools Compared (2026)

The short version: start with laravel/ai for runtime features, add Boost to your development environment, and use specialist packages only when you need something the official SDK doesn't cover.

1. Laravel AI SDK (laravel/ai): The Official Choice

The Laravel AI SDK (laravel/ai) is Laravel's first-party package for working with AI providers. It gives you a consistent API for OpenAI, Anthropic, Gemini, and others, so you can build agents with tools and structured output, generate images, synthesize and transcribe audio, and create embeddings without learning each provider's SDK.

According to the laravel/ai repository, version 1.0.0 shipped on September 23, 2026, and the latest release is v1.3.0 (October 19, 2026). It's MIT-licensed and requires PHP 8.3+ and Laravel 12 or 13.

Install the Laravel AI SDK

composer require laravel/ai
php artisan vendor:publish --provider="Laravel\Ai\AiServiceProvider"
php artisan migrate   # creates the agent conversation tables

Then add your provider keys to .env, such as OPENAI_API_KEY or ANTHROPIC_API_KEY.

Build your first agent

Agents are PHP classes that hold instructions, tools, and an optional output schema. Generate one with Artisan:

php artisan make:agent SupportAssistant

// app/Ai/Agents/SupportAssistant.php
namespace App\Ai\Agents;

use Laravel\Ai\Contracts\Agent;
use Laravel\Ai\Promptable;
use Stringable;

class SupportAssistant implements Agent
{
    use Promptable;

    public function instructions(): Stringable|string
    {
        return 'You answer customer questions about orders. Be brief and friendly.';
    }
}

use App\Ai\Agents\SupportAssistant;
use Laravel\Ai\Enums\Lab;

// Prompt the agent, failing over to Anthropic if OpenAI is rate-limited
$response = (new SupportAssistant)->prompt(
    'Where is my order #1042?',
    provider: [Lab::OpenAI, Lab::Anthropic],
);

return (string) $response;

What the Laravel AI SDK covers

• Agents with conversation memory, tools, sub-agents, middleware, and structured output

• Streaming, broadcasting, and queueing for chat interfaces and background jobs

• MCP tools (with the Laravel MCP package installed) and human tool approval, allowing an agent to request approval before taking risky actions

• Image generation, text-to-speech, and transcription

• Embeddings, reranking, classification, and vector stores for search and RAG

• Provider failover when rate limits or outages occur

• Testing fakes such as SupportAssistant::fake(), so tests never call a real API

Text generation works with OpenAI, Anthropic, Gemini, Azure, Bedrock, Groq, xAI, DeepSeek, Mistral, Ollama, OpenRouter, Cohere, and any OpenAI-compatible endpoint.

For a full walkthrough covering memory, tools, structured output, and streaming, read our Laravel 13 AI SDK guide.

Laravel AI SDK vs. Prism and other packages

If you're starting fresh in 2026, begin with laravel/ai. It's maintained by the Laravel team, follows framework conventions, and covers most common use cases. Keep Prism or openai-php/laravel if they already work well in production or if you need a provider feature the SDK doesn't expose yet.

Our comparison of Laravel AI agent packages covers agent frameworks such as LarAgent and Neuron AI in more detail.

Sara runs a two-person agency in Lisbon. In 2025, she had three client apps using three different AI packages, each with its own configuration and testing setup. After laravel/ai reached 1.0, she started using the SDK for new features. One fake-based testing pattern now covers all three apps, and switching a client from OpenAI to Anthropic became a one-line change.

2. StackMivo: AI App Builder for Laravel

StackMivo works one level above the other Laravel AI tools on this list. Instead of adding AI features to an app, it builds the Laravel app itself. You describe what you need in plain English, and StackMivo generates a full-stack Laravel application with:

• Eloquent models, migrations, and relationships

• Controllers, complete CRUD functionality, and form validation

• Authentication, roles, and admin panels

• REST and GraphQL endpoints

• Your choice of Blade, Livewire, Inertia with React, or Vue

When to use it: Greenfield projects, client MVPs, internal tools, and CRUD-heavy business apps where scaffolding can consume the first few weeks. You own the generated code and can add laravel/ai features or install Laravel Boost afterward, just as you would with any other Laravel project.

Skip the boilerplate, keep the code. Describe your app, and StackMivo generates a Laravel codebase you can extend with the AI SDK. Start building free with StackMivo →

3. Laravel Boost: AI Coding Context for Your Editor

Laravel Boost is the Laravel team's free toolkit for AI coding agents. It runs an MCP server inside your project and installs version-aware guidelines and skills, helping Claude Code, Cursor, Codex, and GitHub Copilot write code that matches your Laravel and package versions.

Highlights in 2026:

• MCP tools to read your database schema, run read-only queries, and check logs and browser errors

• A documentation API with more than 17,000 pieces of Laravel ecosystem knowledge

• Project rules that save your team's conventions in .ai/rules

• Requires PHP 8.2+ and Laravel 11.45.3+, 12.41.1+, or 13; the latest version is 2.10.3

Boost is a development tool, not a runtime package. Your users never interact with it. Our full Laravel Boost guide covers installation and editor setup.

4. Laravel MCP: Let AI Clients Use Your App

Laravel MCP (laravel/mcp, v1.1 as of October 2026) is the official package for building Model Context Protocol servers in Laravel. You define tools, prompts, and resources as PHP classes, and MCP-compatible AI clients can call them. Sanctum or OAuth 2.1 can protect access.

Use it when you want AI assistants to act on your product, such as looking up an order or creating a support ticket. Learn the protocol basics in our Laravel MCP guide.

5. Prism: Fluent Multi-Provider LLM Package

Prism was a go-to Laravel LLM package before the official SDK existed, and it's still actively developed (v0.102.0, October 2026). Among Laravel AI packages, it has one of the most fluent APIs, working across OpenAI, Anthropic, Gemini, Ollama, DeepSeek, Mistral, xAI, and more.

Key strengths:

• One interface, many providers: Switch models without rewriting application logic.

• Structured output: Use schemas instead of parsing free-form text.

• Multimodal input: Work with images, documents, and audio.

• Streaming and tool calling: Build chat and agent workflows.

• Testing helpers: Fake responses in your tests.

Choose Prism if your app already uses it or you prefer its API. For new projects, compare it with laravel/ai first.

6. openai-php/laravel: Direct OpenAI Access

openai-php/laravel is a popular community package created by Nuno Maduro. It wraps the openai-php client with a Laravel facade, configuration file, and testing fakes. It isn't an official Laravel package, but it's widely used and actively maintained (v0.21.0, September 2026).

Use it when you only need OpenAI and want direct access to OpenAI-specific endpoints as soon as they become available.

7. LLPhant: RAG and Agents in PHP

LLPhant is a framework-agnostic PHP library inspired by LangChain and LlamaIndex. It works with Laravel and Symfony and focuses on retrieval-augmented generation (RAG).

• Multi-provider support, including OpenAI, Anthropic, Mistral, and Ollama

• Embeddings and vector store integrations such as Elasticsearch, Milvus, Qdrant, and Redis

• Agent and tool support for multi-step tasks

Choose LLPhant when you need a specific vector database integration or a pipeline you'll reuse outside Laravel.

8. Ollama-Laravel: Run Models Locally

Ollama-Laravel connects Laravel to a local Ollama server, allowing you to run open models on your own hardware. It's useful for privacy-sensitive data, offline development, and experimentation without API bills.

It supports text generation, chat with message history, vision models, embeddings, streaming, and model management. Note that laravel/ai and Prism can also connect to Ollama, so you may not need a separate package unless you want its model-management helpers.

9. Gemini for Laravel: Google's Multimodal Models

Gemini for Laravel is a community package that brings Google's Gemini API to Laravel through a facade. Gemini handles text, code, images, audio, and video, making it a good fit for document and media analysis. The package is actively maintained (2.0.5, October 2026).

As with OpenAI, laravel/ai supports Gemini directly for most tasks. Use this package when you need Gemini-specific endpoints.

10. Laravel Auto Translation: AI-Powered Localization

Laravel Auto Translation, by Vildan Bina, translates your language files using AI and translation APIs.

• Drivers for ChatGPT, Google Translate, and DeepL

• Scans JSON and nested PHP language files

• Preserves placeholders such as :attribute

• Artisan commands to find and translate missing keys

It's a quick win for apps that need to support several languages.

11. TikToken PHP: Count Tokens, Control Costs

AI APIs bill by token, so counting tokens before sending a prompt helps you control costs and stay within context limits. yethee/tiktoken-php is a PHP port of OpenAI's tiktoken tokenizer that works in any Laravel app.

If you use the Laravel AI SDK, also check its built-in usage tracking, which reports token usage per response.

Machine Learning Libraries for Laravel

Not every AI feature needs an LLM. For traditional machine learning in pure PHP, such as classification, regression, clustering, or anomaly detection, use Rubix ML. It offers more than 40 supervised and unsupervised algorithms, preprocessing pipelines, model persistence, and cross-validation, and it's actively maintained (3.3.0, October 2026).

You may still see PHP-ML recommended in older tutorials. Its original GitHub repository is no longer available, and it hasn't kept pace with modern PHP, so Rubix ML is the preferred choice for new projects. For computationally intensive models, many teams run Python services using FastAPI or Flask and call them from Laravel through the HTTP client.

How to Choose the Right Laravel AI Tool

Most Laravel AI tools overlap in some areas, so answer one question first: Who is the AI for?

• Your users (chatbots, summaries, search, image generation): Use the Laravel AI SDK. Add LLPhant for complex RAG or Ollama-Laravel for local models.

• Your developers (writing code faster and more accurately): Install Laravel Boost in your development environment.

• Other AI tools (letting ChatGPT or Claude act on your app): Build a server with Laravel MCP.

• Your next project (getting a working app quickly): Generate it with StackMivo, then layer in the tools above.

Farah, a freelance developer in Toronto, used exactly this approach for a property management client in 2026. She generated the admin app with StackMivo, added a laravel/ai agent to draft replies to tenant messages, and kept Boost running in Cursor for follow-up changes. The client saw a working demo weeks earlier than with her usual timeline.

Best Practices for Using Laravel AI Tools

• Queue slow AI calls. The Laravel AI SDK supports queueing, so users don't have to wait through a 20-second generation.

• Cache repeat results. Store summaries, embeddings, and classifications to reduce API costs.

• Fake AI in tests. Use laravel/ai or Prism fakes so your test suite never makes a paid API call.

• Plan for outages. Configure provider failover for critical features.

• Track usage. Log token usage per feature so costs don't come as a surprise.

• Guard risky tools. Use human approval for agent tools that can delete records, issue refunds, or send emails.

• Keep secrets out of prompts. Never send API keys, passwords, or unnecessary personal data to a model.

Frequently Asked Questions

What is laravel/ai?

laravel/ai is the official Laravel AI SDK. It provides one API for OpenAI, Anthropic, Gemini, and other providers, with agents, tools, structured output, embeddings, images, audio, and testing fakes. Version 1.0 was released on September 23, 2026.

Is the Laravel AI SDK free?

Yes. The package is open source under the MIT license. You still pay AI providers, such as OpenAI or Anthropic, for the API calls your app makes.

Should I use laravel/ai or Prism?

For new projects, start with laravel/ai because it's official and covers most common needs. Prism remains a good choice if your app already uses it or you prefer its fluent API.

Which Laravel AI tools are free?

All the packages on this list are free to install, and the official packages (laravel/ai, Laravel Boost, and Laravel MCP) are MIT-licensed. Usage can still cost money when you call providers such as OpenAI or Anthropic. Ollama-Laravel avoids API fees by running models on your own hardware. StackMivo is free to start.

What's the difference between Laravel Boost and the Laravel AI SDK?

Boost helps you write code by giving AI coding agents context about your project. The AI SDK helps your app use AI at runtime to power features your users interact with.

What is the best AI tool to build a Laravel app?

To generate a complete Laravel application from a description, use an AI app builder such as StackMivo. To help AI coding assistants edit an existing app more effectively, use Laravel Boost.

Conclusion

Laravel AI tools have matured quickly. In 2026, the official laravel/ai SDK is the natural starting point for AI features, Laravel Boost keeps your coding agent on track, and Laravel MCP opens your app to AI clients. Prism, openai-php, LLPhant, and other specialist packages still fill important gaps.

Choose Laravel AI tools based on who the AI serves, queue and cache your calls, and use fakes in your tests. With those practices in place, adding AI to a Laravel app becomes as routine as adding a mailable.

And if you'd rather start with a working app than an empty routes/web.php, let StackMivo generate the foundation.

Build your next Laravel app with StackMivo, free to start →