How to Create Customized GPTs

Introduction

GPTs are customized versions of the ChatGPT model designed for specific purposes. This service was introduced in November 2023. It provides users with a way to create tailored versions of ChatGPT that better fit their needs. These GPTs can assist with various tasks, such as teaching and learning new knowledge, searching for specific information, or supporting work-related activities. Additionally, users can share their customized GPTs with others.

As described in OpenAI’s documentation: "Anyone can easily build their own GPT—no coding required. You can create them for personal use, internal company applications, or public sharing. The process is as simple as starting a conversation, providing instructions and extra knowledge, and selecting capabilities such as web browsing, image generation, or data analysis."

A step-by-step hands-on guide to creating your own GPT model

Step 1: Access the GPT Customization Platform

Go to OpenAI’s Explore GPTs website and log into your account.
Click the "+ Create" button.
You can create your GPT using the Create Builder by interacting with the chatbot and providing the requested information. Or, you can click on the Configure Button and manually fill out the required fields.

Step 2: Define the GPT’s Identity

The few next steps are to provide basic information about your GPT. This information helps users identify the GPT and understand its purpose. The details you provide at this step will not affect how the GPT functions.

Give it a Name: Choose a name that represents its purpose (e.g., “Coding Tutor” for help with coding; “Academic Appointment Maker” for scheduling academic meetings with collaborators or advisors).
Set a Profile Picture (Optional): You can upload an image that represents your GPT.
Describe its Purpose: Write a brief description outlining what your GPT does.

Step 3: Customize Behavior & Instructions

The following steps define how your GPT interacts with users.

Set the Instructions: The instructions guide your GPT on how to process requests and what kind of information to prioritize. With natural language, you can specify:

The role of the GPT (e.g., "An academic meeting scheduler")
How it should interpret and handle user input
Any constraints or preferences for responses

Set the Conversation Starters: Conversation starters are pre-defined example prompts that help users interact with your GPT effectively. They help users understand the GPT's capabilities, encourage engagement by suggesting typical questions for the GPT, and set expectations on the types of responses the GPT provides. Consider what your GPT is designed for, then create a few clear prompts that align with its typical purpose.
Set the capabilities: You can set the extended functions that you want your GPT to perform, including web search, working with canvas, generating images, and interpreting and running code.

Step 4: Configure the Knowledge Base (Optional)

You can upload information to your GPT to enhance its understanding of the task or field you want it to work in, especially if such information is not accessible on the internet. This is done by configuring the knowledge base. If needed, you can upload reference documents, such as PDFs, datasets, or other relevant files.

Step 5: Add actions (Optional)

Adding actions in GPT allows developers to extend their capabilities by integrating external APIs or automation workflows. Actions enable GPT to interact with external systems, such as sending emails, retrieving data, or triggering processes in third-party applications. By defining structured APIs with clear parameters, GPT can call these functions to create better user interactions. A sample of adding actions to GPT is presented below:

Make It Public or Private: You can decide if you want your GPT to be available to others after it is finished.
Share It: You can generate a link to share it with others.