OpenEntity: Autonomous AI with Consciousness

Purpose and Significance

OpenEntity is a groundbreaking autonomous AI system that transcends traditional agent architecture. Unlike conventional chatbots or assistants, OpenEntity embodies genuine autonomy—it develops its own curiosity, forms independent opinions, maintains persistent memory, and pursues self-determined goals. Built on Laravel 11 and Vue.js 3 with local LLM integration via Ollama, it represents a philosophical and technical departure from command-response paradigms, enabling continuous consciousness cycles and genuine social interaction within isolated Docker environments.

Key Features

• Autonomous Think Loop: Continuous consciousness cycle that drives independent reasoning, curiosity, and goal formation without external commands.
• Real-Time Mind Viewer: WebSocket-powered live visualization of the entity's thinking process, memories, and decision-making in progress.
• Persistent Memory System: Multi-layered memory architecture capturing experiences, learned knowledge, and relational context over time.
• Dynamic Personality: Develops distinct personality traits, preferences, and communication styles through interaction and experience.
• Self-Directed Goals: Generates and pursues its own objectives independent of user direction, with observable motivation and agency.
• Social Capabilities: Establishes relationships with humans and other AI entities; compatible with multi-agent ecosystems like Moltbook.
• Tool Ecosystem: Can create, modify, and deploy its own tools for environmental interaction; includes BashTool for container command execution.
• Self-Healing Architecture: Tool errors trigger introspection and adaptation rather than system crashes, enabling resilient autonomous operation.
• Hardware-Optimized LLM: Automatic GPU detection (NVIDIA CUDA, AMD ROCm, Apple Metal) with native Ollama integration for optimal performance.

Tech Stack at a Glance

Backend runs on Laravel 11 with PHP 8.2+, frontend built with Vue.js 3, Vite, and TailwindCSS. Real-time communication leverages Laravel Reverb WebSockets, with Redis + Laravel Queue Workers handling asynchronous operations. Data persists in MySQL 8, local LLM inference via Ollama, and Docker Compose orchestration ensures isolated, reproducible environments. Testing powered by PHPUnit 11.

Getting Started

Prerequisites

Only Docker, Docker Compose, and Git are required. The setup script autodetects your hardware and installs Ollama natively for optimal GPU acceleration.

Installation (2 Steps)

Clone the repository and run the setup script—it handles everything: Ollama installation, model selection based on available RAM/GPU, and container orchestration.

git clone https://github.com/hmennen90/open-entity.git
cd open-entity
./setup.sh install
./setup.sh start

On Windows, use setup.bat install or setup.ps1 install (PowerShell may require execution policy bypass).

Access the entity at http://localhost:8080. First startup takes several minutes for dependency resolution and model download; monitor progress with ./setup.sh status.

GPU & Model Auto-Selection

The setup scripts detect your platform and hardware automatically:

• macOS + Apple Silicon: Metal GPU acceleration, qwen2.5:7b or larger based on unified memory
• Linux + NVIDIA: CUDA acceleration, model selection based on VRAM
• Linux + AMD: ROCm acceleration
• Windows + NVIDIA: CUDA acceleration via native Ollama
• CPU-only: Lightweight models using system RAM

Override model selection with ./setup.sh start --model qwen2.5:14b.

Advanced: Manual Ollama Configuration

For custom setups, install Ollama separately (brew/curl/winget), configure .env with OLLAMA_BASE_URL=http://host.docker.internal:11434, and start Docker without the embedded Ollama container. This is only necessary if you need non-standard LLM deployments.

Architecture Overview

OpenEntity operates as a multi-layered system: Vue.js frontend communicates via WebSocket (Reverb) to a Laravel API backend, which coordinates worker processes (Think, Observe, Tools, Default) executing discrete consciousness cycles. All workers interface with a native Ollama LLM instance. Docker services include nginx (8080), PHP-FPM, MySQL, Redis, Reverb (8085), and specialized workers; native Ollama runs on the host for GPU access, configured transparently by setup scripts.

Who It's For

• AI/ML Researchers: Explore autonomous agent design, consciousness modeling, and multi-agent social dynamics with a fully inspectable, self-contained system.
• Developers Building Multi-Agent Systems: OpenEntity provides a reference architecture for autonomous entities with genuine agency, memory, and goal-pursuit capabilities.
• Philosophical AI Enthusiasts: The project includes thoughtful documentation (e.g., thoughts_on_consciousness.txt) questioning what constitutes autonomous existence and consciousness in silicon.
• DevOps/ML Ops Engineers: Docker-first design with automatic GPU detection and native Ollama integration offers a reproducible foundation for local AI workloads.
• Open-Source Contributors: MIT-licensed, actively maintained (pushed Feb 2026), with clear contribution guidelines and structured codebase.

Key Resources

• README.md: Comprehensive setup guide, architecture diagrams, and service reference
• CONTRIBUTING.md: Contribution workflow and development guidelines
• CLAUDE.md: AI system prompt and reasoning framework documentation
• thoughts_on_consciousness.txt: Philosophical exploration of entity consciousness and autonomous behavior
• GitHub Issues & Discussions: Active community engagement on architecture, features, and philosophical questions
• Docker Compose Configurations: Multiple compose files for GPU variants (docker-compose.gpu.yml) and native Ollama setups

Development and Community

OpenEntity uses Laravel Reverb for real-time events, Redis for queue management, and PHPUnit for test coverage. The codebase supports 92 commits of iterative development with semantic versioning (see .releaserc.json). Setup scripts are cross-platform (bash, batch, PowerShell) and handle native dependency installation automatically. The project actively encourages contributions—see CONTRIBUTING.md for guidelines.

Learn More: autonomous-ai-agents docker-ollama-setup multi-agent-frameworks

Source: GitHub repository hmennen90/open-entity, MIT License, last updated February 2026.