Build LLM Agents on the Fly Without Code With CrewAI

Not a coder? Still interested in building agentic systems to automate business processes? Don’t worry—You can easily build your agentic systems with no-code tools by using either pre-built templates or Crew Studio, which are available on the crewAI platform. In this blog, we’ll explore exciting, new, and lesser-known features of the CrewAI framework by building agents with no-code tools in just a few easy steps.

What is CrewAI?

CrewAI = Crew (A group of people who work together)+ Artificial Intelligence

CrewAI is a popular multi-agent framework that is rapidly gaining a lot of traction in the world of Generative AI. It is designed to create and manage teams of autonomous AI agents that work collaboratively to solve complex problems using large language models (LLMs). These agents can take on specific roles, interact with each other, make independent decisions, and complete multi-step tasks in coordination, much like human teams.

Build Powerful No-Code Agents Instantly with CrewAI Templates

Now, let’s dive into the CrewAI platform. We’ll start by exploring the ready-to-use templates available for creating agents. Below is a step-by-step guide to help you create agents using these templates.

Step 1: Access the CrewAI Platform

Head to the CrewAI platform by clicking here and log in using your email ID. You can choose the free plan and fill in the required details.

Step 2: Find the Right Template

There are various pre-built agentic templates available on CrewAI. To explore all available templates, select Template from the left side of the screen. This will give you access to a range of ready-to-use templates within the platform.

Here are some interesting agent templates you’ll find:

• Sales Offer Generator
• Lead Scoring and Strategy Crew
• Job Change Monitoring Crew
• Meeting Preparation Crew
• Similar Company Finder

Let’s try using the Similar Company Finder template.

Step 3: Crew Deployment Guide

Within the template, you’ll notice API requirements such as Serper API and OpenAPI keys. You’ll need these keys to deploy this agent. SeperAPI key helps in collecting information from web whereas the openAPI key helps in accessing the OpenAI large language models.

You can access your free Serper API keys by signing in with your email ID here: https://serper.dev/.

To learn how to access OpenAI API keys, visit our blog on How to Generate Your Own OpenAI API Key and Add Credits?

At the bottom of the template you’ll see two options in the Similar Company Finder template: Deploy and Download.

Choose Deploy to directly deploy the agent for immediate use.

Once you click Deploy, enter all the required API keys and deploy the Crew template.

Note: This process may take some time to complete.

Your agent is now deployed and ready for direct use according to your needs.

Step 4: Post-Deployment

After deploying your model, go to the Management UI to view all the models deployed from your account. Since we have deployed a Company Finder Agent, you will find this agent on the page.

Step 5: Check the output of the Deployed Agent

To check the output of the deployed agent, click Manage Crew. Here. you will find three sections:

1. Waiting Execution: This is the section where you enter your inputs. Click on Trigger Crew, and type in the details asked, as per your chosen template.

For example:

• Target Company: Samsung
• Our Product: Mobile Phones

Then, select Trigger Crew.

2. Running: Once you’ve specified the inputs, the agent begins processing in this section.

3. Completion: After the agent has processed the inputs, you’ll see the output reflected here. Once an input appears in the completed section, you can select it to view details such as total tokens, prompt tokens, etc.

Here, you will see two options: Output and Tasks.

Click Output to get the final response.

You can click on Tasks to see the list of tasks completed by the agent.

Congratulations! 🎉 You have built your first agent. Feel free to explore other available templates.

But wait—there’s more! What if you have a unique need that these templates don’t fully address? Don’t worry. CrewAI studio offers solutions for that as well.

Build a Custom AI Agent from Scratch using Crew Studio

Note: This platform is still in the beta phase, so there may be minor modifications in the interface.

You can create your own customized agent without code using Crew Studio. Follow these steps to get started:

Step 1: Select Crew Studio

Open Crew Studio from the navigation bar on the left side of the screen.

You will see two options:

• Create LLM connections
• Set Default LLM connection

Complete both these steps before building your agent. Let’s see how that’s done in the next 2 steps.

Step 1A: Create LLM Connections

Start by selecting Create LLM Connection in Crew Studio, or go directly to LLM Connections from the left menu.

Fill in the required details:

• Connection Name: Openai (You can choose another name as well)
• Provider: Select openai as the provider and choose the model as gpt-4o-mini (choose your preferred LLM model based on the requirement)

Note: You can also choose multiple models if needed.

• Environment Variables: Set this as OPENAI_API_KEY

Ensure your environment variables follow this format (e.g:  for Groq: GROQ_API_KEY).

Env-Var-Value: Paste your API key here.

Finally, select Add Connection to complete the setup.

Step 1B: Set Default LLM Connection

Go back to crew sudio and select Set Default LLM Connection in Crew Studio, or go to Settings to configure the default connection.

• Enter all necessary details, including your organization name in Organization Settings
• Set Agent LLM Settings: Select the default language model agents will access.
  • Default LLM Connection: OpenAI
  • Default Model: GPT-4o mini

(Note: The dropdown will display only the models you added in LLM connections. In this example, only OpenAI options are available.)

• Set Crew Studio LLM Settings: Select the default model for creating agents.
  • Default Crew Studio LLM Connection: OpenAI
  • Default Crew Studio LLM Model: GPT-4o mini

Once the above steps are complete, select save default LLM settings.

Step 1C: Set Environment Variable

From the navigation bar go to environment variables and add relevant details

• Key: OPENAI_API_KEY
• Value: Your openai API key

Now, you’re all set! With these prerequisites configured, you’re ready to start building your own agent with Crew studio.

Step 2: Create an Agent

Return to Crew Studio, where you’ll be prompted to describe the type of automation you want to build. Let’s use the prompt “You are a technical blog writer. You write blogs between 1000 to 1500 words based on the technical topic provided.”

You may get a couple of follow-up responses from crewAI, checking about the details of your requirements. Once done you will see an option to generate crew plan.

Upon confirmation, you’ll receive the plan in a tabular format that details each agent’s role, goal, and backstory. It also includes a task breakdown with descriptions, expected outputs, and the agents responsible.

If you need to make changes, you can edit each cell directly by selecting the edit option.

If you’re satisfied with the entire crew plan created for your agent, select Generate Crew.

Step 3: Deploy the Agent

You can view the entire task flow represented in the flow diagram that is generated. Click Deploy Crew to deploy the agent.

Step 4: Check the Output of the Agent

Head to Management UI from the navigation bar and wait for your agent to be built. This may take some time. Once done select Manage Crew. Here also you’ll find sections labelled Waiting Execution, Running, and Completed.

Click Trigger Crew and add the required inputs. You may get different required inputs based on your agent. You could have edited these inputs in the table that was generated earlier in the process.

Then, click on Trigger Crew to start.

Your agent will now appear in the Running section, and once complete, the final output will be available in the Completed section. You can then click on Output to view the response.

Exciting, isn’t it? You’ve just created your own customized agent!

Note: This platform is still in the beta phase, so there may be minor modifications in the interface.

Conclusion

Creating AI agents had never been this easy. With no-code tools, anyone–regardless of technical background can create their own agents. Whether using pre-built templates or designing custom agents from scratch, CrewAI can help you build your agent in just a few steps. So, take the leap, explore what you can build, and start transforming ideas into smart AI agents today!

If you are interested in learning more about AI Agents, checkout our excusive Agentic AI Pioneer Program today!

Frequently Asked Questions

Diksha Kumari

As an Instructional Designer at Analytics Vidhya, Diksha has experience creating dynamic educational content on the latest technologies and trends in data science. With a knack for crafting engaging, cutting-edge content, Diksha empowers learners to navigate and excel in the evolving tech landscape, ensuring educational excellence in this rapidly advancing field.