EvoAgentX Quickstart Guide ¶

This quickstart guide will walk you through the essential steps to set up and start using EvoAgentX. In this tutorial, you'll learn how to:

Configure your API keys to access LLMs
Automatically create and execute workflows

Installation ¶

pip install git+https://github.com/EvoAgentX/EvoAgentX.git

Please refere to Installation Guide for more details about the installation.

API Key & LLM Setup ¶

The first step to execute a workflow in EvoAgentX is configuring your API keys to access LLMs. There are two recommended methods to configure your API keys:

Method 1: Set Environment Variables in the Terminal ¶

This method sets the API key directly in your system environment.

For Linux/macOS:

export OPENAI_API_KEY=<your-openai-api-key>

For Windows Command Prompt:

set OPENAI_API_KEY=<your-openai-api-key>

For Windows PowerShell:

$env:OPENAI_API_KEY="<your-openai-api-key>" # " is required

Once set, you can access the key in your Python code with:

import os
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")

Method 2: Use a `.env` File ¶

You can also store your API key in a .env file inside the root folder of your project.

Create a file named .env with the following content:

OPENAI_API_KEY=<your-openai-api-key>

Then, in your Python code, you can load the environment settings using python-dotenv:

from dotenv import load_dotenv 
import os 

load_dotenv() # Loads environment variables from .env file
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")

🔐 Tip: Never commit your .env file to public platform (e.g., GitHub). Add it to .gitignore.

Configure and Use the LLM in EvoAgentX ¶

Once your API key is configured, you can initialize and use the LLM as follows:

from evoagentx.models import OpenAILLMConfig, OpenAILLM

# Load the API key from environment
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")

# Define LLM configuration
openai_config = OpenAILLMConfig(
    model="gpt-4o-mini",       # Specify the model name
    openai_key=OPENAI_API_KEY, # Pass the key directly
    stream=True,               # Enable streaming response
    output_response=True       # Print response to stdout
)

# Initialize the language model
llm = OpenAILLM(config=openai_config)

# Generate a response from the LLM
response = llm.generate(prompt="What is Agentic Workflow?")

You can find more details about supported LLM types and their parameters in the LLM module guide.

Automatic WorkFlow Generation and Execution ¶

Once your API key and language model are configured, you can automatically generate and execute agentic workflows in EvoAgentX. This section walks you through the core steps: generating a workflow from a goal, instantiating agents, and running the workflow to get results.

First, let's import the necessary modules:

from evoagentx.workflow import WorkFlowGenerator, WorkFlowGraph, WorkFlow
from evoagentx.agents import AgentManager

Step 1: Generate WorkFlow and Agents ¶

Use the WorkFlowGenerator to automatically create a workflow based on a natural language goal:

goal = "Generate html code for the Tetris game that can be played in the browser."
wf_generator = WorkFlowGenerator(llm=llm)
workflow_graph: WorkFlowGraph = wf_generator.generate_workflow(goal=goal)

WorkFlowGraph is a data structure that stores the overall workflow plan — including task nodes and their relationships — but does not yet include executable agents.

You can optionally visualize or save the generated workflow:

# Visualize the workflow structure (optional)
workflow_graph.display()

# Save the workflow to a JSON file (optional)
workflow_graph.save_module("/path/to/save/workflow_demo.json")

We provide an example generated workflow here. You can reload the saved workflow:

workflow_graph = WorkFlowGraph.from_file("/path/to/save/workflow_demo.json")

Step 2: Create and Manage Executable Agents ¶

Use AgentManager to instantiate and manage agents based on the workflow graph:

agent_manager = AgentManager()
agent_manager.add_agents_from_workflow(workflow_graph, llm_config=openai_config)

Step 3: Execute the Workflow ¶

Once agents are ready, you can create a WorkFlow instance and run it:

workflow = WorkFlow(graph=workflow_graph, agent_manager=agent_manager, llm=llm)
output = workflow.execute()
print(output)

For a complete working example, check out the full workflow demo.