Creating an AI Agent to Write Weblog Posts with CrewAI

I like writing. You might discover that in the event you observe me or my weblog. For that purpose, I’m continually producing new content material and speaking about Knowledge Science and Synthetic Intelligence.

I found this ardour a few years in the past after I was simply beginning my path in Knowledge Science, studying and evolving my expertise. At the moment, I heard some extra skilled professionals within the space saying {that a} good research method was training new expertise and writing about it someplace, educating no matter you discovered.

As well as, I had simply moved to the US, and no person knew me right here. So I needed to begin someplace, creating my skilled picture on this aggressive market. I keep in mind I talked to my cousin, who’s additionally within the Tech trade, and he informed me: write weblog posts about your experiences. Inform individuals what you might be doing. And so I did. 

And I by no means stopped.

Quick ahead to 2025, now I’ve virtually 200 printed articles, a lot of them with In the direction of Knowledge Science, a broadcast Guide, and a great viewers. 

Writing helped me a lot within the Knowledge Science space.

Most lately, one in every of my pursuits has been the wonderful Pure Language Processing and Massive Language Fashions topics. Studying about how these trendy fashions work is fascinating. 

That curiosity led me to experiment with Agentic Ai as nicely. So, I discovered about CrewAI, a simple and open-source package deal that helps us construct AI brokers in a enjoyable and straightforward method, with little code. I made a decision to check it by making a crew of brokers to write down a weblog publish, after which see how that goes.

On this publish, we are going to discover ways to create these brokers and make them work collectively to provide a easy weblog publish.

Let’s do this.

What’s a Crew?

A crew of AI brokers is a mix of two or extra brokers, every of them performing a process in direction of a last objective.

On this case research, we are going to create a crew that can work collectively to provide a small weblog publish a couple of given subject that we are going to present.

The stream works like this:

1. We select a given subject for the brokers to write down about.
2. As soon as the crew is began, it would go to the information base, learn a few of my beforehand written articles, and attempt to mimic my writing model. Then, it generates a set of tips and passes it to the following agent.
3. Subsequent, the Planner agent takes over and searches the Web searching for good content material in regards to the subject. It creates a plan of content material and sends it to the following agent.
4. The Author agent receives the writing plan and executes it in accordance with the context and knowledge acquired.
5. Lastly, the content material is handed to the final agent, the Editor, who opinions the content material and returns the ultimate doc because the output.

Within the following part, we are going to see how this may be created.

Code

CrewAI is a superb Python package deal as a result of it simplifies the code for us. So, let’s start by putting in the 2 wanted packages.

pip set up crewai crewai-tools

Subsequent, if you need, you possibly can observe the directions on their Quickstart web page and have a full undertaking construction created for you with simply a few instructions on a terminal. Principally, it would set up some dependencies, generate the folder construction steered for CrewAI tasks, in addition to generate some .yaml and .py information. 

I personally choose to create these myself, however it’s as much as you. The web page is listed within the References part.

Folder Construction

So, right here we go.

We are going to create these folders:

And these information:

• Within the config folder: create the information brokers.yaml and duties.yaml
• Within the information folder, that’s the place I’ll add the information with my writing model.
• Within the undertaking root: create crew.py and fundamental.py.

Be certain to create the folders with the names talked about, as CrewAI seems to be for brokers and duties contained in the config folder and for the information base inside a information folder.

Subsequent, allow us to set our brokers. 

Brokers

The brokers are composed of:

• Title of the agent: writer_style
• Function: LLMs are good function gamers, so right here you possibly can inform them which function to play.
• Objective: inform the mannequin what the objective of that agent is.
• Backstory: Describe the story behind this agent, who it’s, what it does. 

writer_style:
  function: >
    Writing Fashion Analyst
  objective: >
    Completely learn the information base and study the traits of the crew, 
    resembling tone, model, vocabulary, temper, and grammar.
  backstory: >
    You might be an skilled ghost author who can mimic any writing model.
    You understand how to determine the tone and elegance of the unique author and mimic 
    their writing model.
    Your work is the idea for the Content material Author to write down an article on this subject.

I received’t bore you with all of the brokers created for this crew. I consider you bought the thought. It’s a set of prompts explaining to every agent what they’ll do. All of the brokers directions are saved within the brokers.yaml file.

Consider it as in the event you had been a supervisor hiring individuals to create a group. Take into consideration what sorts of execs you would want, and what expertise are wanted.

We want 4 professionals who will work in direction of the ultimate objective of manufacturing written content material: (1) a Author Stylist, (2) a Planner, (3) a Author, and (4) an Editor. 

If you wish to see the setup for them, simply examine the total code within the GitHub repository.

Duties

Now, again to the analogy of the supervisor hiring individuals, as soon as we “employed” our total crew, it’s time to separate the duties. We all know that we need to produce a weblog publish, we’ve got 4 brokers, however what every of them will do.

Properly, that might be configured within the file duties.yaml.

As an example, let me present you the code for the Author agent. As soon as once more, these are the elements wanted for the immediate:

• Title of the duty: write
• Description: The outline is like telling the skilled the way you need that process to be carried out, similar to we might inform a brand new rent the best way to carry out their new job. Give exact directions to get the perfect outcome potential.
• Anticipated output: That is how we need to see the output. Discover that I give directions like the dimensions of the weblog publish, the amount of paragraphs, and different info that helps my agent to offer me the anticipated output. 
• Agent to carry out it: Right here, we’re indicating the agent who will carry out this process, utilizing the identical title set within the brokers.yaml file.
• Output file: Now all the time relevant, but when so, that is the argument to make use of. We requested for a markdown file as output.

write:
  description: >
    1. Use the content material plan to craft a compelling weblog publish on {subject}.
    2. Incorporate search engine optimisation key phrases naturally.
    3. Sections/Subtitles are correctly named in a fascinating method. Be certain 
    so as to add Introduction, Downside Assertion, Code, Earlier than You Go, References.
    4. Add a summarizing conclusion - That is the "Earlier than You Go" part.
    5. Proofread for grammatical errors and alignment with the author's model.
    6. Use analogies to make the article extra partaking and complicated ideas simpler
    to know.
  expected_output: >
    A well-written weblog publish in markdown format, prepared for publication.
    The article should be inside a 7 to 12 minutes learn.
    Every part should have not less than 3 paragraphs.
    When writing code, you'll write a snippet of code and clarify what it does. 
    Watch out to not add an enormous snippet at a time. Break it in cheap chunks.
    Within the examples, create a pattern dataset for the code.
    Within the Earlier than You Go part, you'll write a conclusion that's partaking
    and factually correct.
  agent: content_writer
  output_file: blog_post.md

After the brokers and duties are outlined, it’s time to create our crew stream.

Coding the Crew

Now we are going to create the file crew.py, the place we are going to translate the beforehand introduced stream to Python code.

We start by importing the wanted modules.

#Imports
import os
from crewai import Agent, Activity, Course of, Crew, LLM
from crewai.undertaking import CrewBase, agent, crew, process
from crewai.information.supply.pdf_knowledge_source import PDFKnowledgeSource
from crewai_tools import SerperDevTool

We are going to use the essential Agent, Activity, Crew, Course of and LLM to create our stream. PDFKnowledgeSource will assist the primary agent studying my writing model, and SerperDevTool is the software to look the web. For that one, ensure that to get your API key at https://serper.dev/signup.

A finest follow in software program improvement is to maintain your API keys and configuration settings separate out of your code. We’ll use a .env file for this, offering a safe place to retailer these values. Right here’s the command to load them into our surroundings.

from dotenv import load_dotenv
load_dotenv()

Then, we are going to use the PDFKnowledgeSource to point out the Crew the place to seek for the author’s model. By default, that software seems to be on the information folder of your undertaking, thus the significance of the title being the identical.

# Information sources

pdfs = PDFKnowledgeSource(
    file_paths=['article1.pdf',
                'article2.pdf',
                'article3.pdf'
                ]
)

Now we are able to arrange the LLM we need to use for the Crew. It may be any of them. I examined a bunch of them, and people I appreciated essentially the most had been qwen-qwq-32b and gpt-4o. When you select OpenAI’s, you’ll need an API Key as nicely. For Qwen-QWQ, simply uncomment the code and remark out the OpenAI’s traces.. You want an API key from Groq. 

# LLMs

llm = LLM(
    # mannequin="groq/qwen-qwq-32b",
    # api_key= os.environ.get("GROQ_API_KEY"),
    mannequin= "gpt-4o",
    api_key= os.environ.get("OPENAI_API_KEY"),
    temperature=0.4
)

Now we’ve got to create a Crew Base, exhibiting the place CrewAI can discover the brokers and duties configuration information.

# Creating the crew: base exhibits the place the brokers and duties are outlined

@CrewBase
class BlogWriter():
    """Crew to write down a weblog publish"""
    agents_config = "config/brokers.yaml"
    tasks_config = "config/duties.yaml"

Brokers Features

And we’re able to create the code for every agent. They’re composed of a decorator @agent to point out that the next operate is an agent. We then use the category Agent and point out the title of the agent within the config file, the extent of verbosity, being 1 low, 2 excessive. You may as well use a Boolean worth, resembling true or false.

Lastly, we specify if the agent makes use of any software, and what mannequin it would use.

# Configuring the brokers
    @agent
    def writer_style(self) -> Agent:
        return Agent(
                config=self.agents_config['writer_style'],
                verbose=1,
                knowledge_sources=[pdfs]
                )

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

    @agent
    def content_writer(self) -> Agent:
        return Agent(
        config=self.agents_config['content_writer'],
        verbose=1
        )

    @agent
    def editor(self) -> Agent:
        return Agent(
        config=self.agents_config['editor'],
        verbose=1
        )

Duties Features

The subsequent step is creating the duties. Equally to the brokers, we are going to create a operate and beautify it with @process. We use the category Activity to inherit CrewAI’s functionalities after which level to the duty for use from our duties.yaml file for use for every process created. If any output file is predicted, use the output_file argument.

# Configuring the duties    

    @process
    def model(self) -> Activity:
        return Activity(
        config=self.tasks_config['mystyle'],
        )

    @process
    def plan(self) -> Activity:
        return Activity(
        config=self.tasks_config['plan'],
        )

    @process
    def write(self) -> Activity:
        return Activity(
        config=self.tasks_config['write'],
        output_file='output/blog_post.md' # That is the file that might be include the ultimate weblog publish.
        )

    @process
    def edit(self) -> Activity:
        return Activity(
        config=self.tasks_config['edit']
        )

Crew

To attach all the things collectively, we now create a operate and beautify it with the @crew decorator. That operate will line up the brokers and the duties within the order to be carried out, for the reason that course of chosen right here is the only: sequential. In different phrases, all the things runs in sequence, from begin to end.

@crew

    def crew(self) -> Crew:
        """Creates the Weblog Submit crew"""

        return Crew(
            brokers= [self.writer_style(), self.planner(), self.content_writer(), self.editor(), self.illustrator()],
            duties= [self.style(), self.plan(), self.write(), self.edit(), self.illustrate()],
            course of=Course of.sequential,
            verbose=True
        )

Operating the Crew

Operating the crew may be very easy. We create the fundamental.py file and import the Crew Base BlogWriter created. Then we simply use the features crew().kickoff(inputs) to run it, passing a dictionary with the inputs for use to generate the weblog publish.

# Script to run the weblog author undertaking

# Warning management
import warnings
warnings.filterwarnings('ignore')
from crew import BlogWriter


def write_blog_post(subject: str):
    # Instantiate the crew
    my_writer = BlogWriter()
    # Run
    outcome = (my_writer
              .crew()
              .kickoff(inputs = {
                  'subject': subject
                  })
    )

    return outcome

if __name__ == "__main__":

    write_blog_post("Worth Optimization with Python")

There it’s. The result’s a pleasant weblog publish created by the LLM. See under.

That’s so good!

Earlier than You Go

Earlier than you go, know that this weblog publish was 100% created by me. This crew I created was an experiment I needed to do to study extra about the best way to create AI brokers and make them work collectively. And, like I mentioned, I like writing, so that is one thing I’d have the ability to learn and assess the standard.

My opinion is that this crew nonetheless didn’t do an excellent job. They had been in a position to full the duties efficiently, however they gave me a really shallow publish and code. I’d not publish this, however not less than it could possibly be a begin, possibly. 

From right here, I encourage you to study extra about CrewAI. I took their free course the place João de Moura, the creator of the package deal, exhibits us the best way to create totally different sorts of crews. It’s actually attention-grabbing.

GitHub Repository

https://github.com/gurezende/Crew_Writer

About Me

If you wish to study extra about my work, or observe my weblog (it’s actually me!), listed below are my contacts and portfolio.

https://gustavorsantos.me

References

[Quickstart CrewAI](https://docs.crewai.com/quickstart)

[CrewAI Documentation](https://docs.crewai.com/introduction)

[GROQ](https://groq.com/)

[OpenAI](https://openai.com)

[CrewAI Free Course](https://study.crewai.com/)