agenticos 0.0.3.156160

Tapping into the power of AI agents

AgenticOS

AgenticOS is a flexible and modular framework designed to help developers deploy and manage AI-driven workflows using agent-based systems. The framework is designed to be LLM-agnostic and core-tech switchable in the future, allowing developers to integrate their preferred agent systems seamlessly.

The goal is to provide a framework that integrates easily into your existing or planned infrastructure, giving you the flexibility to focus on AI task execution without complex setups.

Geting started

Installing the prerequisites

pipx

This is currently a recomended way to install comamnds that are available accros your systems like build backends or other code generators. Please follow instructions for your operating system: https://pipx.pypa.io/stable/installation/

hatch

This build backend is required by crew AI. You can install it with pipx:

pipx install hatch

For other installation options please check the official instructions: https://python-poetry.org/docs/#ci-recommendations

CrewAI

Please install CrewAI using the pipx:

pipx install crewai

:warning: Please make sure that you update CrewAI before every use.

Creating a new project

1. Create a new CrewAI crew. If you are installing agenticos alongside exising CrewAI project please go to point 3.

crewai create crew new_crew

2. Generate venv and install dependencies:

cd new_crew
hatch shell

3. Install agenticos package using poetry:

pip install agenticos

4. Add agenticos to pyproject.toml. Use the following command to find out which version do you need.

pip freeze | grep agenticos

5. Generate agentic scaffold files:

agentic init

It will generate following files in your project:

[app_root]
|-src
  |-agentic
    |-agentic_node.py
    |-agentic_config.py

:warning: This configuration should work with out-of the box crewai project. If you made any changes to project structure you will have to update agentic_config.py file.

Running

Currentic agentic node may be run only as a standalone HTTP server. Production fastapi mode:

agentic run

or dev fastapi mode:

agentic run --dev

Config via ENV VARS

Order of priority CLI ARGS > ENV_VARS > DEFAULTS. Current env vars with defaults:

    AUTH_TOKEN: str = ""
    REGISTRY_URL: str = "ws://localhost:8080"
    NODE_MODE: str = "httpserver"
    HTTP_PORT: int = 8000
    AGENTIC_CONFIG_PATH: str = "src/agentic/agentic_config.py"
    HTTP_HEALTHCHECK : bool = False