Requirements for Running Agent Zero: Complete Docker Setup Guide

Docker is the only mandatory requirement for running Agent Zero, which executes inside a pre-built container supporting Windows, macOS, and Linux.

Agent Zero is an open-source AI agent framework maintained by agent0ai/agent-zero. Understanding the requirements for running Agent Zero ensures a smooth deployment, whether you use the recommended containerized approach or configure a local Python development environment.

Mandatory Requirements for Running Agent Zero

Docker Installation

Docker is the sole mandatory component for running Agent Zero. The entire framework executes inside a pre-built Docker container, guaranteeing a consistent environment across all platforms. You need either Docker Desktop (for Windows and macOS) or Docker Engine (for Linux) installed and running on your host machine.

According to the installation documentation in docs/setup/installation.md, no other system dependencies are required when using the Docker image, as all Python packages and system libraries are bundled inside the container.

Operating System Compatibility

Agent Zero supports any operating system capable of running Docker:

• Windows: Windows 10 or Windows 11
• macOS: macOS 12 (Monterey) or later, including Apple Silicon (M1/M2/M3) support
• Linux: Any recent distribution capable of running Docker CE (Community Edition) or Docker Desktop for Linux

Optional Local Development Requirements

Python Dependencies

If you plan to run the codebase directly—for extension development, debugging, or testing—you must install the Python dependencies listed in requirements.txt. The main runtime dependencies include:

• Core frameworks: flask[async] (3.0.3), langchain-core (0.3.49), langchain-community (0.3.19)
• AI/ML libraries: browser-use (0.5.11), faiss-cpu (1.11.0), sentence-transformers (3.0.1), openai-whisper (20250625)
• Infrastructure: docker (7.1.0), paramiko (3.5.0), playwright (1.52.0)
• Data processing: unstructured[all-docs] (0.16.23), pypdf (6.0.0), markdown (3.7)

These exact versions are specified in requirements.txt at the repository root.

Development and Testing Tools

For running the test suite or contributing code, install the additional packages from requirements.dev.txt:

• pytest
• pytest-asyncio
• pytest-mock

Configuration and Environment Setup

Environment Variables

Agent Zero supports automated configuration via environment variables using the A0_SET_<setting> prefix. You can define these in a .env file at the project root to pre-configure model providers, API keys, and memory options without manually entering them through the web UI.

As documented in docs/setup/installation.md, this approach allows you to start the container with all settings pre-loaded, streamlining deployment in automated environments.

Hardware Specifications

While Agent Zero runs on modest hardware, performance scales with available resources:

• CPU: Any modern multi-core processor; GPU is optional and only required for local model inference (e.g., Ollama)
• RAM: Minimum 2 GB; 8 GB or more recommended for larger language models or heavy memory usage
• Storage: Several gigabytes required for the Docker image, cached models, and persistent data stored in the /a0/usr folder

Quick Start: Running Agent Zero with Docker

To launch Agent Zero with the mandatory Docker requirement, execute the following commands:

docker pull agent0ai/agent-zero
docker run -p 50001:80 \
  -v "$HOME/agent-zero-data:/a0/usr" \
  agent0ai/agent-zero

This command pulls the pre-built image containing all Python dependencies and maps port 50001 on your host to port 80 in the container. The volume mount persists user data to $HOME/agent-zero-data on your host machine.

Key Source Files for Setup Reference

When configuring your environment, consult these specific files in the agent0ai/agent-zero repository:

File	Purpose
README.md	High-level overview, quick-start snippet, and feature list
docs/setup/installation.md	Detailed Docker installation steps and environment variable configuration
requirements.txt	Exact Python package versions for local development
DockerfileLocal	Build instructions and system library dependencies
initialize.py	Entry point that loads settings and starts the Flask/WSGI server

Summary

• Docker is mandatory for running Agent Zero; the framework executes entirely within a pre-built container.
• Operating system support includes Windows 10/11, macOS 12+, and Linux distributions capable of running Docker.
• Local development requires Python 3.x and dependencies from requirements.txt, including Flask, LangChain, and browser automation libraries.
• Hardware minimums are 2 GB RAM and a modern multi-core CPU, with 8 GB+ recommended for production workloads.
• Configuration can be automated via A0_SET_* environment variables in a .env file.

Frequently Asked Questions

Is Docker mandatory for running Agent Zero?

Yes. Docker is the only mandatory requirement for running Agent Zero. The framework is designed to run inside a pre-built Docker container that bundles all Python dependencies, system libraries, and runtime configurations, ensuring consistent behavior across Windows, macOS, and Linux.

What operating systems support Agent Zero?

Agent Zero supports any operating system capable of running Docker, specifically Windows 10 or 11, macOS 12 (Monterey) and later (including Apple Silicon), and recent Linux distributions. As long as you can install Docker Desktop or Docker Engine, you can run Agent Zero.

Can I run Agent Zero without Python installed locally?

Yes. If you use the Docker image, you do not need Python installed on your host machine. The container includes all necessary Python packages listed in requirements.txt. You only need Python locally if you plan to run the source code directly for development or debugging purposes.

How much RAM do I need to run Agent Zero?

The minimum RAM requirement is 2 GB, but 8 GB or more is recommended for optimal performance, especially when using larger language models or heavy memory features. If you plan to run local models via Ollama, additional RAM or GPU resources will improve inference speed.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}