Tutorial: Build Your First AI Agent (2 Free Paths)

What you will build

You will build a simple AI agent that can:

Take a task from the user
Think about it step by step
Produce structured output
(Optional) keep simple memory

You will build the same idea in two ways:

No-code — Flowise + Ollama
Code — Python + Ollama

Both paths lead to a working first agent.

Copy Commands

Use these command snippets directly.

# Ollama model
ollama run llama3.1

# Check Node.js
node -v
npm -v

# Start Flowise
npx flowise start

# Python deps
pip install requests

PART A — No-Code AI Agent (Flowise + Ollama)

Best for beginners, non-technical users, and fast understanding.

STEP 1 — Install Ollama (Local AI Model)

Go to https://ollama.com, install it, then run:

ollama run llama3.1

STEP 2 — Install Node.js

Download and install the LTS version from https://nodejs.org, then verify:

node -v
npm -v

STEP 3 — Install Flowise

npx flowise start

Open: http://localhost:3000

STEP 4 — Create Your First Flow

In Flowise, create a new chat flow and add:

User Input
Prompt Template
ChatOllama
Output

STEP 5 — Add Chat Model Node

Use:

Base URL: http://localhost:11434
Model: llama3.1

STEP 6 — Add Prompt Template

You are an AI Agent.
Break tasks into steps and respond clearly.

Task: {input}

STEP 7 — Connect the System

User Input → Prompt Template → ChatOllama → Output

STEP 8 — Test Your Agent

Plan a simple website structure for a coffee shop

STEP 9 — Improve Structure

You are an AI Agent.

Follow this structure:
1. Understand the task
2. Break it into steps
3. Provide a final structured answer

Task: {input}

STEP 10 — Add System Thinking Instruction

Always think step-by-step before answering.

STEP 11 — Save Your Flow

Save as My First AI Agent.

STEP 12 — What You Built

You built a controlled reasoning pipeline, not just a chatbot.

PART B — Code-Based AI Agent (Python + Ollama)

Best for developers who want to understand the internals.

STEP 1 — Install Python

Install from https://python.org, then verify:

python --version

STEP 2 — Install Ollama

ollama run llama3.1

STEP 3 — Install Python Dependency

pip install requests

STEP 4 — Create Project File

Create agent.py.

STEP 5 — Basic AI Connection

import requests

MODEL = "llama3.1"
URL = "http://localhost:11434/api/generate"

def ask_ai(prompt):
    response = requests.post(
        URL,
        json={"model": MODEL, "prompt": prompt, "stream": False},
        timeout=60,
    )
    return response.json()["response"]

STEP 6 — Create Your Agent Logic

def agent(task):
    prompt = f"""
You are an AI Agent.

Step 1: Understand the task
Step 2: Break it into steps
Step 3: Provide a structured solution

Task: {task}
"""
    return ask_ai(prompt)

STEP 7 — Run Your Agent

result = agent("Design a simple login system")
print(result)

STEP 8 — Add Simple Memory (Optional)

memory = []

def agent(task):
    context = "\n".join(memory[-5:])
    prompt = f"""
You are an AI Agent.

Memory:
{context}

Task: {task}
"""
    result = ask_ai(prompt)
    memory.append(task)
    memory.append(result)
    return result

STEP 9 — Test Multiple Calls

print(agent("Create a to-do app"))
print(agent("Now improve it"))

STEP 10 — What Changed?

The agent now maintains context across requests.

STEP 11 — Why This Is an Agent

It receives input, runs logic, maintains memory, and returns structured output.

STEP 12 — Final insight

An AI agent is not only the model; it is the system around the model.

Summary

You built:

A no-code agent (Flowise + Ollama)
A code agent (Python + Ollama)

Same principle:

AI ≠ model — AI = system (the orchestration, prompts, tools, and memory around the model).

Tutorial - Build Your First AI Agent (2 Free Paths)