CrewAI Studio Quickstart

Quick Start Guide to Use CrewAI Studio (GUI):

1. Launch CrewAI Studio in your browser using https://public_ip_of_vm/crewai-studio

2. Configure Environment:

Before starting, create or edit the .env file in the project folder to add your API keys for supported LLM providers (e.g., OpenAI). The .env_example file is provided as a template. To edit the .env file

• Connect to your instance via SSH terminal and navigate to CrewAI-Studio directory using -

cd /home/ubuntu/setup/CrewAI-Studio/

• Here you will find .env_example , create a .env file from it and move it to /app directory.

cp .env_example  .env

mv  .env app/

• Now edit .env file to add your LLM Provider details. This step is crucial as the crew won’t run without it.

cd app/

vi .env

• Press “i” to enable insert mode , uncomment the line of LLM provide of your choice and provide its value. Save the file by pressing ESC key followed by :wq . Now your LLM provider is set, go back to CrewAI Studio in your browser.

3. Add Knowledge Sources & Tools:

• Use the interface to upload knowledge bases or connect to APIs/tools your agents may require.

• Custom tools and enhanced features like web scraping or code interpreters can also be added via GUI.

4. Create Agents for your crews:

Agents are specialized AI entity designed to perform specific roles within a collaborative framework of Crews. Provide Role , Goal and Backstory for your agent, select the LLM provider you set in your .env file as explained above in step 2. You can create AI agents visually, specifying roles, tasks, and tool integrations without coding.

5. Create new task from task option. Provide Description, expected output, choose agent to whom you want to assign this task and provide other details. Save the task.

6. Create a New Crew:

In the GUI, click on the option to create a new crew/project. Crea is a collection of agents working together to complete a mission. Give a name to your crew, provide whether you want to run the tasks sequentially or hierarchically (parallelly) , select agents, task , llm provider and other configuration on this page. Save the changes.

7. Run Your Crew:

Once you are ready with your crew, navigate to Kickoff page, select your crew to run and hit Run Crew button. You can expand the Console Output to see the output of the kickoff process.

6. View Results & History Results from each run are displayed in the interface. You can also Print the result.

You can browse previous runs and export crews or results as single-page apps or code if desired.

7. Manage & Export Export your crews as simple, shareable Single page apps/JSON.

Save configurations and manage different AI agent projects from one central GUI.

CrewAI Studio makes it easy to build, manage, and run multi-agent AI projects without writing code while supporting advanced customizations via config and tools when needed.

To get started with CrewAI Platform please visit CrewAI Quickstart Guide

To Learn more about JupyterHub, multiuser environment and installing new python packages on this VM visit JupterHub Guide

For more details, please visit Official Documentation page