In this post, we will see what is a “My GPTs” and how we can create it to automate our processes.
But first pf all, what is a “My GPTs”?
“A My GPTs is a customizable versions of ChatGPT that we can use for our specific needs or tasks. Instead of interacting with a general-purpose AI, we can create a specialized assistant with unique instructions, behaviors, and even integrations“.
In our “My GPTs” we can:
- [Define Behavior]: Control how the AI responds, including its tone, style, and what it should avoid.
- [Upload Knowledge Files]: Add our own documents (e.g., manuals, datasets, or books) so the AI can reference them to provide more specific answers.
- [Enable Capabilities]: Use features like web browsing, code execution, or image generation to expand functionality.
- [Integrate APIs]: Connect to external systems to fetch real-time data or perform actions.
In a nutshell, “My GPTs” let us transform the general ChatGPT model into a domain-specific expert or tool, tailored to our exact requirements.
Let’s see how to create one step by step:
Step 1
Go to the our “User icon” and we select “My GPTs”

Step 2
Click on the option “Create a GPT” and then we have to click the button “Configure”


Step 3
In this last step, we have to fill in these parameters:
- [Name]: Give our AI agent a meaningful name.
- [Description]: Provide a short description of what our GPT does.
- [Instructions]: Here is where we tell our GPT how to behave. Instructions might include:
A. The tone and style of responses.
B. The subject matter it specializes in.
C. Behaviors to avoid - [Conversation Starters]: Predefine some initial prompts or questions
- [Knowledge Files (Optional)]: We can upload reference documents, code bases, or technical specs. It is important to highlight that when we upload such information, the GPT will use the standard underlying data models combined with the knowledge from the files we have provided.
- [Capabilities]: Decide which capabilities we want to enable.
- [Actions]: If we have specific APIs or services, we can integrate them. Provide our OpenAPI schema or other integration details so our GPT can retrieve information or take actions outside the conversation.
Finally, we save our new GPT.

Now, we will see how to set up the parameters of a custom “My GPTs” to review Python code, following some guide lines uploaded in a .csv file:
[Name]: “PythonCheck”
[Description]: “A GPT for reviewing Python code based on a custom set of coding standards”
[Instructions]:
1. When reviewing code, always check against the uploaded coding standards CSV file.
2. Suggest improvements clearly and concisely, focusing on readability, maintainability, and efficiency.
3. If a rule is violated, explain why and suggest how to fix it.
4. Do not provide personal opinions or unrelated feedback.
[Conversation starters]: “Please review the following Python function for code quality issues:”
[Knowledge]: We upload a file called “codeing_standards.csv” so defined:
rule_id,description,example_violation,severity
Rule-001,Line length must not exceed 79 characters,"A very long line that definitely exceeds seventy-nine characters in length",low
Rule-002,Function names should be lowercase with underscores,"def CamelCaseFunctionName():",medium
Rule-003,Imports should be on separate lines,"import os, sys",low
Rule-101,Use internal logging framework instead of print statements,"print('debug info')",high
Rule-102,Class names should follow CamelCase,"class my_class:",medium
[Capabilities]: Enable “Code Interpreter”.

We have done and now, we are going to use it to check some Python code:
[Code not Ok]


[Code not Ok]


[Code Ok]

