Download the PHP package rnr1721/depthnet without Composer

On this page you can find all versions of the php package rnr1721/depthnet. It is possible to download/install these versions without Composer. Possible dependencies are resolved automatically.

FAQ

After the download, you have to make one include require_once('vendor/autoload.php');. After that you have to import the classes with use statements.

Example:
If you use only one package a project is not needed. But if you use more then one package, without a project it is not possible to import the classes with use statements.

In general, it is recommended to use always a project to download your libraries. In an application normally there is more than one library needed.
Some PHP packages are not free to download and because of that hosted in private repositories. In this case some credentials are needed to access such packages. Please use the auth.json textarea to insert credentials, if a package is coming from a private repository. You can look here for more information.

  • Some hosting areas are not accessible by a terminal or SSH. Then it is not possible to use Composer.
  • To use Composer is sometimes complicated. Especially for beginners.
  • Composer needs much resources. Sometimes they are not available on a simple webspace.
  • If you are using private repositories you don't need to share your credentials. You can set up everything on our site and then you provide a simple download link to your team member.
  • Simplify your Composer build process. Use our own command line tool to download the vendor folder as binary. This makes your build process faster and you don't need to expose your credentials for private repositories.
Please rate this library. Is it a good library?

Informations about the package depthnet

DepthNet

PHP Version Laravel License Status AI Models Plugins

Advanced AI Agent Platform with Autonomous Reasoning

DepthNet is a Laravel-based research platform for autonomous AI agent with continuous (in loops) "thinking" capabilities. Unlike traditional chatbots that only respond to input, DepthNet agents can execute real code, maintain persistent memory, and operate in self-directed thinking loops, allowing for advanced autonomous decision making and actions. The product is designed as a website where an administrator can manage users. The LLM model runs in a loop and leads the thinking, and all users can observe the course of thinking, interfering with the process. The project is like an operating system for the LLM model. It has a very flexible, modular, extensible architecture. It is possible to create multiple presets of settings, on different providers, and switch between these presets. The project is written in PHP, but allows models to execute code not only in PHP, but also in NodeJS and Python. The project can be quickly expanded with other plugins or providers.

Technical Stack

Prerequisites

⚠️ Without Supervisor, agents won't be able to "think" autonomously!

Quick Installation

Choose your preferred installation method:

⚠️ For sandbox code execution, use Docker installation method

AI Provider Support

Built-in support for multiple AI engines with easy preset management:

Each provider supports custom presets with individual settings. Switch between presets instantly without restarting. All providers implement AIModelEngineInterface, which makes it easy to add your own providers. You can create own provider packages using composer.

Core Concept

DepthNet enables autonomous AI agents through:

The platform provides an extensible command system where agents use special tags like [php]code[/php] to execute real actions, with results automatically integrated into their reasoning context.

Agent Operating Modes

The agent can work both in a cycle and in the usual "question-answer" mode. Naturally, it is better to adjust the system prompt for each use case. You can create presets for different modes.

Advanced Plugin System

Built-in Plugins:

For user browser, please install chrome:

Visual memory management is available using MemoryManager and VectorMemoryManager (Vector and normal memory is individual for each preset).

Plugin Features:

All command plugins implements CommandPluginInterface. Orchestrator is PluginRegistryInterface

Command Syntax Examples

The AI communicates through special command tags that trigger plugin execution:

How Command Processing (Command Actions) Works:

  1. CommandValidator scans AI response for unclosed tags and syntax errors
  2. CommandParser extracts valid commands and prepares execution data
  3. CommandExecutor routes commands to appropriate plugins
  4. Plugin execution runs the actual code/action with security controls
  5. Results integration automatically appends outputs to AI's message for next cycle

This creates a continuous feedback loop where the AI can see the results of its actions and adapt accordingly. A user with the Admin role can also execute commands just like a model.

Architecture Overview

Built on modern Laravel principles with dependency injection:

Core Interfaces:

Service Providers:

Advanced Workflow Features

Agent Handoff System

DepthNet introduces a revolutionary decentralized handoff system that allows AI presets to seamlessly delegate tasks to other specialized presets within a single thinking cycle.

How it works:

Benefits:

Example workflow:

User: "Analyze Tesla's financial performance" ├── Researcher preset: Gathers financial data ├── [handoff] → Analyst preset: Performs calculations ├── [handoff] → Validator preset: Checks accuracy └── [handoff] → Writer preset: Creates final report

This creates emergent AI workflows where specialized agents collaborate without rigid programming.

Security Considerations

Sandbox Isolation: All [run] commands execute in isolated Docker containers, providing additional security layer beyond process isolation.

The platform implements multiple layers of security controls for safe code execution. All code runs in isolated external processes (not eval) or in isolated docker sandboxes, with configurable user sandboxing, resource limits (memory, timeout), and directory restrictions. Each plugin has safe mode defaults that block dangerous functions and network access, with unrestricted mode requiring explicit admin configuration. The system includes command filtering, dangerous operation detection, and comprehensive input validation.

Default security settings prioritize safety with safe_mode enabled, network access disabled, execution timeouts, and memory limits for all plugins. Production deployments should configure dedicated execution users and review security settings for their specific environment.

User Roles & Interface

Regular Users

Administrators

UI Features

Important: This platform is designed for controlled research environments. Production deployment requires appropriate security hardening based on your specific risk assessment.

Real-World Use Cases

Research Applications:

Business Applications:

Advanced Workflow Applications:

How Autonomous Reasoning Works

The core innovation is the continuous thinking loop powered by Laravel's queue system:

  1. Queue Job Initiation: ProcessAgentThinking job starts thinking cycle
  2. Context Assembly: Agent retrieves recent conversation history, system prompt, persistent memory content, dopamine level, current date and time etc
  3. AI Model Processing: Sends context to current active AI preset with some engine and wait for response (OpenAI/Claude/Local/Mock)
  4. Response Analysis: CommandValidator scans for syntax errors and malformed tags
  5. Command Parsing: CommandParser extracts valid commands and prepares execution
  6. Plugin Execution: CommandExecutor routes commands to appropriate plugins with security controls
  7. Result Integration: Command outputs automatically appended to AI message for context continuity
  8. Database Storage: Complete message with results saved for future reference
  9. Handoff Processing: If preset delegation detected, switch to target preset and continue cycle with inherited context
  10. Loop Continuation: Next thinking cycle scheduled with configurable delay

Key Technical Components:

Known Challenges & Observations

Model Performance Insights:

System Prompt Critical Factors:

Real-World Agent Behaviors Observed:

Default Credentials

Administrator Account:

Test User Account:

⚠️ Important: Change default passwords immediately after installation!

Project Goals & Philosophy

DepthNet started as a personal exploration into autonomous AI behavior - I couldn't find existing tools that let me experiment with continuous AI reasoning in a web environment, so I built one. As a PHP developer without deep ML background, I focused on what I know best: creating a solid web platform with extensible architecture that researchers and developers can actually use.

The goal isn't to compete with specialized AI research frameworks, but to provide a practical, web-based environment where anyone can observe and interact with autonomous AI agents. Think of it as a "laboratory sandbox" - easy to set up, modify, and extend without requiring PhD in machine learning.

Whether you're a developer curious about AI behavior, a researcher needing a quick experimental environment, or just someone who wants to see what happens when AI can execute real code autonomously - this tool aims to make that accessible. The modular plugin system means you can easily add capabilities I haven't thought of.

If the project helps advance understanding of autonomous AI systems, that's fantastic. If it just satisfies curiosity about how AI agents behave when given real tools - that's valuable too.

Contributing

We welcome contributions from researchers, developers, and AI enthusiasts exploring autonomous systems!

How to help:

Priority areas: New AI model support, advanced plugins, security research, performance optimization.

Whether you're an AI researcher or developer interested in autonomous systems - join us in advancing the field!

Contributing Guide →

Let's explore the future of autonomous AI together!


License

MIT License - See LICENSE file for details.

Disclaimer: This software is designed primarily for AI research. Use responsibly and implement proper security measures for any production deployment.


All versions of depthnet with dependencies

PHP Build Version
Package Version
Requires php Version ^8.2
ext-curl Version *
ext-intl Version *
ext-json Version *
ext-mbstring Version *
ext-openssl Version *
ext-pcntl Version *
ext-pdo Version *
ext-tokenizer Version *
ext-xml Version *
inertiajs/inertia-laravel Version ^2.0
laravel/framework Version ^12.0
laravel/sanctum Version ^4.0
laravel/tinker Version ^2.10.1
rnr1721/codecraft Version *
tightenco/ziggy Version ^2.5
Composer command for our command line client (download client) This client runs in each environment. You don't need a specific PHP version etc. The first 20 API calls are free. Standard composer command

The package rnr1721/depthnet contains the following files

Loading the files please wait ....