Your First OpenClaw Agent — Hello World to Production

What Are OpenClaw Agents?

OpenClaw is an open-source framework for building AI agents — autonomous programs that can reason, plan, and take actions to accomplish goals. Unlike simple chatbots that respond to a single prompt, agents maintain context, use tools, and make decisions across multiple steps.

Think of an agent as an AI-powered assistant that can:
Break complex tasks into smaller steps
Use external tools (APIs, databases, file systems)
Remember context across interactions
Make decisions based on intermediate results

Setting Up Your Environment

First, let's install OpenClaw and its dependencies:

python

# Create a virtual environment
python -m venv openclaw-env
source openclaw-env/bin/activate  # On Windows: openclaw-env\Scripts\activate

# Install OpenClaw
pip install openclaw

# Verify the installation
python -c "import openclaw; print(openclaw.__version__)"

You'll also need an API key for your chosen LLM provider. Set it as an environment variable:

python

import os
os.environ["OPENAI_API_KEY"] = "your-key-here"
# Or use Anthropic, Google, or any supported provider

Building Your First Agent

Let's build a simple question-answering agent. This agent will take a user's question, think through its response, and provide a well-structured answer.

python

from openclaw import Agent, AgentConfig

# Configure the agent
config = AgentConfig(
    name="my-first-agent",
    model="gpt-4o",  # Or "claude-sonnet-4-20250514", "gemini-pro", etc.
    system_prompt="""You are a helpful research assistant.
    When answering questions:
    1. Think step by step
    2. Cite your reasoning
    3. Be concise but thorough""",
    temperature=0.7,
    max_tokens=1024,
)

# Create the agent
agent = Agent(config)

# Run the agent with a question
response = agent.run("What are the key differences between SQL and NoSQL databases?")
print(response.output)

This creates a basic agent with a system prompt that guides its behavior. The AgentConfig lets you specify the model, temperature, and token limits.

Adding Memory

Agents become much more powerful when they can remember previous interactions:

python

from openclaw import Agent, AgentConfig, ConversationMemory

config = AgentConfig(
    name="memory-agent",
    model="gpt-4o",
    system_prompt="You are a helpful assistant that remembers our conversation.",
    memory=ConversationMemory(max_turns=10),
)

agent = Agent(config)

# First interaction
r1 = agent.run("My name is Alex and I'm learning Python.")
print(r1.output)

# Second interaction — the agent remembers!
r2 = agent.run("What's my name and what am I learning?")
print(r2.output)  # Will correctly recall "Alex" and "Python"

Testing and Iterating

Good agents are built through iteration. Here's a simple testing pattern:

python

test_cases = [
    {"input": "What is 2+2?", "expected_contains": "4"},
    {"input": "Who wrote Romeo and Juliet?", "expected_contains": "Shakespeare"},
    {"input": "What's the capital of France?", "expected_contains": "Paris"},
]

for test in test_cases:
    response = agent.run(test["input"])
    passed = test["expected_contains"].lower() in response.output.lower()
    status = "PASS" if passed else "FAIL"
    print(f"[{status}] {test['input']}")

Key Takeaways

Agents are more than chatbots — they reason, plan, and use tools
Start simple — a basic agent with a good system prompt goes a long way
Add memory to enable multi-turn conversations
Test systematically — define expected behaviors and validate them
Iterate on your system prompt — it's the most impactful parameter

In the next course, you'll learn prompt engineering techniques that make your agents dramatically more reliable.

Ready for more? 🚀

Subscribe for $29/mo to unlock all courses + get 40% off certifications. New content added every week.

Subscribe Now — $29/mo

🔒 Secure payment via Stripe💰 30-day money-back guarantee🚫 Cancel anytime