CrewAI: Generating a Report.

TL;DR.

The post provides a step-by-step guide on deploying and utilising CrewAI agents by following a part of Tyler Reed's YouTube tutorial. This article emphasises learning by doing, documenting the process, and sharing my insights. I highlight the importance of experimenting and creating posts that enhance my understanding of CrewAI's functionalities.

Attributions:

https://www.youtube.com/watch?v=ONKOXwucLvE from Tyler AI ↗.

---

An Introduction.

Although I will be deploying CrewAI agents, my main objective is to show the results of converting hands-on processes into documentation:

The purpose of this post is to show how practical experiences are captured as documented events.

---

The Big Picture.

This is the first project (in a series of projects) where I cover part of a YouTube video from Tyler Reed.

Tyler AI

If you are new to CrewAI (like me), then I STRONGLY suggest you do EXACTLY what I did: Follow Tyler’s instructions, make detailed notes about his process, turn your notes into a blog post, and publish your results. You will learn by doing and you will end up with a post that you can reference. Copying and pasting my post is not a learning process. You will ultimately cheat yourself, and fail to understand how CrewAI works.

Here is a link to Tyler’s video:

https://www.youtube.com/watch?v=ONKOXwucLvE&t=0 ↗.

---

Prerequisites.

• A Debian-based Linux distro (I use Ubuntu),

• CrewAI.

---

Updating my Base System.

• From the (base) terminal, I update my (base) system:

sudo apt clean && \
sudo apt update && \
sudo apt dist-upgrade -y && \
sudo apt --fix-broken install && \
sudo apt autoclean && \
sudo apt autoremove -y

NOTE: The Ollama LLM manager is already installed on my (base) system.

---

What is the Report Project?

The Report Project shows how to deploy existing tools for CrewAI agents to use. A list of existing tools can be found within the CrewAI documentation ↗.

---

Creating the Report Project.

• From the terminal, I navigate to my projects directory:

cd ~/AI

• I start a new project:

crewai create crew report

• From the first list of options, I choose the ollama option.

• From the next list of options, I choose the ollama/llama3.1 option.

Opening the Report Project in VS Code.

• I open the Report Project in VS Code:

code ./report

Editing the .env File.

• I open the .env file and replace the contents with the following:

OPENAI_API_BASE=https://openrouter.ai/api/v1 # openrouter url here
OPENAI_MODEL_NAME=openrouter/google/gemini-2.0-flash-001 # openrouter/path/model_name here
OPENROUTER_API_KEY=your_api_key # API key here

NOTE: By default, these settings provide this project with access to the OpenRouter.ai API.

Editing the main.py File.

• Under the src/report directory, I open the main.py file.

• I replace the contents of the main.py file with the following:

#!/usr/bin/env python
import sys
import warnings

from datetime import datetime

from crew import Report

warnings.filterwarnings("ignore", category=SyntaxWarning, module="pysbd")

# This main file is intended to be a way for you to run your
# crew locally, so refrain from adding unnecessary logic into this file.
# Replace with inputs you want to test with, it will automatically
# interpolate any tasks and agents information

def run():
    """
    Run the crew.
    """
    inputs = {
        'topic': 'AI LLMs',
        'current_year': str(datetime.now().year)
    }

    try:
        Report().crew().kickoff(inputs=inputs)
    except Exception as e:
        raise Exception(f"An error occurred while running the crew: {e}")

run()

What changed:

• Near the top, I changed from report.crew import Report to from crew import Report,

• I deleted the train(), replay(), and test() functions, and

• At the bottom of the file, I added the run() function.

Editing the crew.py File.

• Under the src/report directory, I open the crew.py file.

• I replace the contents of the crew.py file with the following:

from crewai import Agent, Crew, Process, Task, LLM
from crewai.project import CrewBase, agent, crew, task
from dotenv import load_dotenv

load_dotenv()

# If you want to run a snippet of code before or after the crew starts,
# you can use the @before_kickoff and @after_kickoff decorators
# https://docs.crewai.com/concepts/crews#example-crew-class-with-decorators

@CrewBase
class Report():
    """Report crew"""

    # Learn more about YAML configuration files here:
    # Agents: https://docs.crewai.com/concepts/agents#yaml-configuration-recommended
    # Tasks: https://docs.crewai.com/concepts/tasks#yaml-configuration-recommended
    agents_config = 'config/agents.yaml'
    tasks_config = 'config/tasks.yaml'

    ollama_llm = LLM(
        model = 'ollama/deepseek-r1:14b',
        base_url = 'http://localhost:11434'
    )

    # If you would like to add tools to your agents, you can learn more about it here:
    # https://docs.crewai.com/concepts/agents#agent-tools
    @agent
    def researcher(self) -> Agent:
        return Agent(
            config=self.agents_config['researcher'],
            verbose=True,
            llm = self.ollama_llm
        )

    @agent
    def reporting_analyst(self) -> Agent:
        return Agent(
            config=self.agents_config['reporting_analyst'],
            verbose=True,
            llm = self.ollama_llm
        )

    # To learn more about structured task outputs,
    # task dependencies, and task callbacks, check out the documentation:
    # https://docs.crewai.com/concepts/tasks#overview-of-a-task
    @task
    def research_task(self) -> Task:
        return Task(
            config=self.tasks_config['research_task'],
        )

    @task
    def reporting_task(self) -> Task:
        return Task(
            config=self.tasks_config['reporting_task'],
            output_file='report.md'
        )

    @crew
    def crew(self) -> Crew:
        """Creates the Report crew"""
        # To learn how to add knowledge sources to your crew, check out the documentation:
        # https://docs.crewai.com/concepts/knowledge#what-is-knowledge

        return Crew(
            agents=self.agents, # Automatically created by the @agent decorator
            tasks=self.tasks, # Automatically created by the @task decorator
            process=Process.sequential,
            verbose=True,
            # process=Process.hierarchical, # In case you wanna use that instead https://docs.crewai.com/how-to/Hierarchical/
        )

What changed:

• At the top of the page and the end of the line from crewai import Agent, Crew, Process, Task I added LLM,

• Near the top, I added from dotenv import load_dotenv and load_dotenv(), and

• Within the Report() class, after tasks_config = 'config/tasks.yaml' I added the following:

    ollama_llm = LLM(
            model = 'ollama/deepseek-r1:14b',
            base_url = 'http://localhost:11434'
        )
  

• Within the researcher() and reporting_analyst() agents, I added llm = self.ollama_llm to override the OpenRouter.ai settings in the .env file.

Editing the agents.yaml File.

• Under the src/report/config directory, I open the agents.yaml file.

• I replace the contents of the agents.yaml file with the following:

researcher:
  role: >
    {topic} Senior Data Researcher
  goal: >
    Uncover cutting-edge developments in {topic}
  backstory: >
    You are a seasoned researcher with a knack for uncovering the latest
    developments in {topic}. Known for your ability to find the most relevant
    information and present it in a clear and concise manner.

reporting_analyst:
  role: >
    {topic} Reporting Analyst
  goal: >
    Create detailed reports based on {topic} data analysis and research findings
  backstory: >
    You are a meticulous analyst with a keen eye for detail. You are known for
    your ability to turn complex data into clear and concise reports, making
    it easy for others to understand and act on the information you provide.

What changed:

• No changes made.

Editing the tasks.yaml File.

• Under the src/report/config directory, I open the tasks.py file.

• I replace the contents of the tasks.py file with the following:

research_task:
  description: >
    Conduct a thorough research about {topic}
    Make sure you find any interesting and relevant information given
    the current year is {current_year}.
  expected_output: >
    A list with 10 bullet points of the most relevant information about {topic}
  agent: researcher

reporting_task:
  description: >
    Review the context you got and expand each topic into a full section for a report.
    Make sure the report is detailed and contains any and all relevant information.
  expected_output: >
    A fully fledged report with the main topics, each with a full section of information.
    Formatted as markdown without '```'
  agent: reporting_analyst

What changed:

• No changes made.

---

The Results.

This CrewAI project provides a comprehensive guide to deploying and utilising CrewAI agents effectively. By watching Tyler Reed's video, following his instructions, and documenting the process, I gained a deeper understanding of CrewAI's functionalities. This hands-on approach not only enhanced my learning but also resulted in a valuable reference post. As I embark on this journey, I must experiment, document my findings, and share my insights with the community.

---

In Conclusion.

I discovered how to deploy, and utilise, CrewAI agents. I followed along with (some of) Tyler Reed's video tutorial, learned by doing, and enhanced my understanding of CrewAI's functionalities. This process is perfect for me as I look to document my journey while sharing my insights.

Until next time: Be safe, be kind, be awesome.

---

Hash Tags.

#CrewAI #AI #AI-Projects #AI-Tools #AI-Research #AI-Community

#AI-Integration #Tech-Tutorial #Tech-Blog #Open-Source

#VS-Code #Ubuntu #Learning-By-Doing #Practical-AI