neuron-ai

Create Full-Featured Agentic Applications in PHP

[!IMPORTANT] Get early access to new features, exclusive tutorials, and expert tips for building AI agents in PHP. Join a community of PHP developers pioneering the future of AI development. Subscribe to the newsletter

Before moving on, support the Neuron AI community giving a GitHub star ⭐️. Thank you!

What is Neuron

Neuron is a PHP framework for creating and orchestrating AI Agents. It allows you to integrate AI entities into your PHP applications with a powerful and flexible architecture. We provide tools for the entire agentic application development lifecycle, from LLM interfaces, data loading, multi-agent orchestration, to monitoring and debugging. In addition, we provide tutorials and other educational content to help you get started using AI Agents in your projects.

Video Tutorial

Why Neuron

Your next application will be agentic. A growing share of new software is no longer a web application with AI features added along the way, but an application born agentic, where the agent is the architecture itself, driving how the system reasons, acts, and talks to the user interface. Building this kind of application requires a specific set of foundations: event-driven workflows with checkpointing, human-in-the-loop, interruption, multi-agent orchestration, streaming, and agentic UI integration like AG-UI and the Vercel AI SDK, MCP connectors, and asynchronous execution.

Each one is a chapter of the documentation: Workflow, Human in the loop, Streaming & UI protocols, MCP, Async.

There is also no second framework waiting for you when the project grows. The same Workflow that runs your first agent in the getting started guide runs a multi-agent system with state, loops, and human approvals in production. What you learn on day one is what you ship in future projects.

A Vertical Ecosystem

Neuron is the vertical ecosystem for agentic applications development in PHP. Around the framework there is a registry of extensions, tools, and technologies designed specifically for agentic applications, and a growing number of companies building on the same architecture instead of assembling their own from scattered parts.

For a software house or a product company this is a place to be recognized as a specialist rather than one more team claiming AI experience. For a company that needs an agentic foundation it can commit to for years, it means standardizing on an architecture whose whole direction is this space, not a general-purpose library where AI is a side feature. It also means hiring becomes a solvable problem: as the community grows, so does the number of developers who already know how to design, test, and run agentic applications on this architecture, so the people you bring in tomorrow speak the same language as the system you are building today.

Our Official Channels

Follow the Neuron channels to stay up to date:

• Documentation: https://docs.neuron-ai.dev/
• Newsletter: https://neuron-ai.dev
• Inspector: https://inspector.dev
• E-Book: https://www.amazon.it/dp/B0F1YX8KJB
• Linkedin: https://www.linkedin.com/company/neuron-ai-php-framework
• X: https://x.com/neuronai_php
• Instagram: https://www.instagram.com/neuronai_p

Start With One Prompt

Copy & paste this initial prompt to tell your coding agent how to install and configure Neuron AI in your project.

You are going to set up Neuron AI, a PHP framework for building agentic applications,
in this project, and you must follow three steps in order.

- First: install the framework: check composer.json, and if neuron-core/neuron-ai is not
already required, run `composer require neuron-core/neuron-ai`;

- Second: install the agent skills that ship
with the package by running `npx --yes skills add ./vendor/neuron-core/neuron-ai/skills -y` from
the project root. Add this command to `post-update-cmd` in the composer.json file, so they
stay updated whenever Neuron is updated through Composer. From then on, load the relevant
skill before you write any Neuron code instead of relying on what you remember about
the framework;

- Third: work out what kind of app this is. Look at composer.json
and the project structure to see whether it is an application based on Laravel, Symfony, or another framework,
then activate the matching skill: neuron-laravel-integration or neuron-symfony-integration.
If Neuron is already part of the app go through the skill's foundations checklist
item by item. Report what differs from the recommended setup;

If Neuron is not in the app yet, follow the same checklist to do the first setup.
If the app uses neither framework, start from a plain PHP agent using
the neuron-agent skill. Finish with a short summary of what you installed
and what you found.

Requirements

• PHP: ^8.1

How To

• Getting Started
• AI Providers
• Tools & Toolkits
• MCP Connector
• Structured Output
• RAG
• Workflow
• AI Assisted Development
• Security Vulnerabilities
• Official Documentation

1) Install

composer require neuron-core/neuron-ai

2) Create an Agent

Neuron provides you with the Agent class you can extend to inherit the main features of the framework and create fully functional agents. This class automatically manages some advanced mechanisms for you, such as memory, tools, up to RAG (Retrieval Augmented Generation). You can go deeper into these aspects below.

Let's create an Agent with the command below:

vendor/bin/neuron make:agent DataAnalystAgent

<?php

namespace App\Neuron;

use NeuronAI\Agent\Agent;
use NeuronAI\Chat\Messages\SystemMessage;
use NeuronAI\Providers\AIProviderInterface;
use NeuronAI\Providers\Anthropic\Anthropic;

class DataAnalystAgent extends Agent
{
    protected function provider(): AIProviderInterface
    {
        return new Anthropic(
            key: 'ANTHROPIC_API_KEY',
            model: 'ANTHROPIC_MODEL',
        );
    }

    protected function instructions(): SystemMessage
    {
        return new SystemMessage(
            "You are a data analyst expert in creating reports from SQL databases."
        );
    }
}

3) Talk to the Agent

Bind the agent to a conversation thread, then send it a message and get a response:

$agent = DataAnalystAgent::make()->setThreadId('user-42');


$response = $agent->chat(
    new UserMessage("Hi, I'm Valerio. Who are you?")
)->getMessage();
echo $response->getContent();
// I'm a data analyst. How can I help you today?


$response = $agent->chat(
    new UserMessage("Do you remember my name?")
)->getMessage();
echo $response->getContent();
// Your name is Valerio, as you said in your introduction.

As you can see in the example above, the Agent has memory of the ongoing conversation. Learn more about memory in the documentation.

Supported LLM Providers

With Neuron, you can switch between LLM providers with just one line of code, without any impact on your agent implementation. Supported providers:

• Anthropic
• OpenAI
• OpenAI Responses API
• OpenAI on Azure
• OpenAILike (OpenAI compatible APIs)
• Ollama
• Gemini
• Gemini Vertex
• Mistral
• HuggingFace
• Deepseek
• Grok
• AWS Bedrock Runtime
• Cohere
• ZAI
• Alibaba DashScope
• Neuron Router

Tools & Toolkits

Make your agent able to perform concrete tasks, like reading from a database, by adding tools or toolkits (collections of tools).

<?php

namespace App\Neuron;

use NeuronAI\Agent\Agent;
use NeuronAI\Chat\Messages\SystemMessage;
use NeuronAI\Providers\AIProviderInterface;
use NeuronAI\Providers\Anthropic\Anthropic;
use NeuronAI\Tools\Toolkits\MySQL\MySQLSchemaTool;
use NeuronAI\Tools\Toolkits\MySQL\MySQLSelectTool;
use NeuronAI\Tools\Toolkits\MySQL\MySQLToolkit;

class DataAnalystAgent extends Agent
{
    protected function provider(): AIProviderInterface
    {
        return new Anthropic(
            key: 'ANTHROPIC_API_KEY',
            model: 'ANTHROPIC_MODEL',
        );
    }

    protected function instructions(): SystemMessage
    {
        return new SystemMessage(
            "You are a data analyst expert in creating reports from SQL databases."
        );
    }

    protected function tools(): array
    {
        return [
            MySQLToolkit::make(
                \DB::connection()->getPdo()
            )->only([MySQLSchemaTool::class, MySQLSelectTool::class]),
        ];
    }
}

The select tool runs each query in a read-only transaction, so the database refuses any write. It can still read everything the connection's user can read. For production, consider giving the toolkit its own connection, with a user that can only SELECT the tables the agent needs and a statement timeout.

Ask the agent something about your database:

$response = DataAnalystAgent::make()->setThreadId('demo')->chat(
    new UserMessage("How many orders we received today?")
)->getMessage();

echo $response->getContent();

Learn more about Tools in the documentation.

MCP Connector

Instead of implementing tools manually, you can connect tools exposed by an MCP server with the McpConnector component:

<?php

namespace App\Neuron;

use NeuronAI\Agent\Agent;
use NeuronAI\MCP\McpConnector;

class DataAnalystAgent extends Agent
{
    ...

    protected function tools(): array
    {
        return [
            // Connect to an MCP server
            ...McpConnector::make([
                'command' => 'npx',
                'args' => ['-y', '@modelcontextprotocol/server-everything'],
            ])->tools(),
        ];
    }
}

Learn more about MCP connector in the documentation.

Structured Output

There are scenarios where you need Agents to understand natural language, but output in a structured format, like business processes automation, data extraction, etc. to use the output with other downstream systems.

use App\Neuron\MyAgent;
use NeuronAI\Chat\Messages\UserMessage;
use NeuronAI\StructuredOutput\SchemaProperty;

/*
 * Define the output structure as a PHP class.
 */
class Person
{
    #[SchemaProperty(
        description: 'The user name',
        required: true
    )]
    public string $name;

    #[SchemaProperty(
        description: 'What the user love to eat'
    )]
    public string $preference;
}

/*
 * Talk to the agent requiring the structured output
 */
$person = MyAgent::make()->setThreadId('demo')->structured(
    new UserMessage("I'm John and I like pizza!"),
    Person::class
);

echo $person->name ' like '.$person->preference;
// John like pizza

Learn more about Structured Output on the documentation.

RAG

To create a RAG, you need to attach some additional components other than the AI provider, such as a vector store, and an embeddings provider.

Let's create a RAG with the command below:

vendor/bin/neuron make:rag MyChatBot

Here is an example of a RAG implementation:

<?php

namespace App\Neuron;

use NeuronAI\Providers\AIProviderInterface;
use NeuronAI\Providers\Anthropic\Anthropic;
use NeuronAI\RAG\Embeddings\EmbeddingsProviderInterface;
use NeuronAI\RAG\Embeddings\VoyageEmbeddingProvider;
use NeuronAI\RAG\RAG;
use NeuronAI\RAG\VectorStore\PineconeVectorStore;
use NeuronAI\RAG\VectorStore\VectorStoreInterface;

class MyChatBot extends RAG
{
    protected function provider(): AIProviderInterface
    {
        return new Anthropic(
            key: 'ANTHROPIC_API_KEY',
            model: 'ANTHROPIC_MODEL',
        );
    }

    protected function embeddings(): EmbeddingsProviderInterface
    {
        return new VoyageEmbeddingProvider(
            key: 'VOYAGE_API_KEY',
            model: 'VOYAGE_MODEL'
        );
    }

    protected function vectorStore(): VectorStoreInterface
    {
        return new PineconeVectorStore(
            key: 'PINECONE_API_KEY',
            indexUrl: 'PINECONE_INDEX_URL'
        );
    }
}

Learn more about RAG in the documentation.

Workflow

Think of a Workflow as a smart flowchart for your AI applications. The idea behind Workflow is to allow developers to use all the Neuron components like AI providers, embeddings, data loaders, chat history, vector store, etc., as standalone components to create totally customized agentic systems.

Agent and RAG classes represent a ready to use implementation of the most common patterns when it comes to retrieval use cases, or tool calls, structured output, etc. Workflow allows you to program your agentic system completely from scratch. Agent and RAG can be used inside a Workflow to complete tasks as any other component if you need their built-in capabilities.

Neuron Workflow supports a robust human-in-the-loop pattern, enabling human intervention at any point in an automated process. This is especially useful in large language model (LLM)-driven applications where model output may require validation, correction, or additional context to complete the task.

Learn more about Workflow on the documentation.

AI Assisted Development

When working with AI coding assistants like Claude Code, Opencode, Cursor, or other similar tools, you have two options to give the AI deep context about Neuron components. This leads to more accurate code suggestions, better understanding of component APIs, and fewer hallucinations when generating Neuron code:

• Skills
• MCP server

Security Vulnerabilities

If you discover a security vulnerability within Neuron, please send an e-mail to the Inspector team via [email protected]. All security vulnerabilities will be promptly addressed.

Official documentation

Go to the official documentation

Development

To run the full test suite locally, use Docker Compose to spin up all the services required by the CI environment:

# Start all services
docker compose up -d

Run the test suite:

# Install dependencies and run tests
docker compose run --rm php composer update --prefer-stable
docker compose run --rm php vendor/bin/phpunit

To stop all services:

docker compose down