Custom ChatGPT App Development: Step-by-Step Tutorial

by | Feb 14, 2024

Imagine being able to create your own Custom ChatGPT App that can Verify A Phone Number. You don’t need any coding experience. By following a few simple steps, you can design a custom Numlookup AI. This AI can validate phone numbers and perform other tasks, similar to a phone number reverse lookup API or a cell phone carrier lookup API.

Making a custom GPT for ChatGPT is an exciting project. You can integrate external APIs to access data and Verify A Phone Number. In this tutorial, we’ll use the Numverify phone validator API to show how to create a custom GPT for ChatGPT. Numverify provides strong, real-time phone number validation and information services. It is perfect for demonstrating how external APIs can improve ChatGPT.

By integrating a phone number validation API into your custom GPT, you can create a GPT that can scan documents for phone numbers and validate them. It can also generate and return files with validated numbers, making processes easier and improving data accuracy. This can improve customer interactions and automate data management tasks, opening up many new possibilities. Let’s explore all the steps.

Tutorial for GTP builder to have own custom gpt chatbot for number validation training data and custom chatbot user interface

Step 1: Access Your Paid Numverify Account

As a paid user of Numverify, you have access to many great features for phone number validation and information retrieval. In this simple guide, we will show you how to integrate the Numverify API into a custom GPT. This will let you chat with an AI that not only talks smartly but also provides real-time data and phone number analysis.

Step 2: Run a Numverify Query

Go to your Numverify account dashboard and click ‘Run’ to run the phone number validation query.

numverify for ai models and advanced natural language processing to automate repetitive tasks for private data and customer satisfaction

A browser window will open and you should get a result similar to this:

Followed by its JSON result

JSON result for validate to create gpts for business processes and ai customization

Keep that browser tab open, because later, you will be asked to paste the URL and the JSON result into ChatGPT so that it can assist you in creating a GPT schema, where your newly created GPT will be able to access the Numverify API.

Step 3: Creating a Custom GPT for the Chat GPT Website

Go to the ChatGPT website. You may already know about OpenAI’s GPT store and its different models.

Now, let’s create your own Custom GPT. This one will use an external API, like the Numverify API.

First, go to the ‘Explore GPTs’ section on the OpenAI platform. You’ll see many pre-built GPT models for different uses. This will help you understand how these models work and give you ideas for your own.


Once you’re in the ‘Explore GPTs’ section, look for the option to ‘Create‘ a new GPT model. This is where you’ll begin the process of building your custom model. Clicking on ‘Create’ will open up a new interface where you can start inputting the specifics for your custom GPT.

gpt create button

Define Your Custom GPT’s Parameters

In this step, you will set up the key parts of your custom GPT model. This includes naming it, choosing its language settings, and deciding other basic features. These choices are important because they create the main structure of your custom model.

Here’s how to start:

  1. Add an Image: Click the “+” button to upload an image.
  2. Name and Describe: Give your GPT a name and write a brief description.
  3. Basic Instructions: Write some simple instructions for your GPT to follow.
  4. Conversation Starter: Include a message to start the conversation.

The screenshot below shows how easy it is to begin. You can add more details to the instructions later.

Simple Example

Your custom GPT needs only a basic outline with a few main points in the instructions to start working.

custom GPT skeleton outline instructions

Customize the Conversation Flow (Advanced example): In the advanced example below, the wording was generated by using the ‘Create’ tab where you can fine-tune your requirements using a chat.

Integrating the Numverify API isn’t just about technical connectivity; it’s also about ensuring that the conversation flow is natural and intuitive. Customize the ‘Instructions’ section for how your GPT model will use the Numverify data in its responses. This could involve setting up specific triggers for the API interaction or scripting responses that effectively incorporate the data retrieved from Numverify.

With the basic structure of your GPT model set up, it’s now time to integrate the Numverify API. This will involve inputting the API key and setting up the necessary endpoints for your GPT model to interact with. You’ll need to configure how your model sends requests to Numverify and how it processes the responses within the conversational context.

To get started with integrating the Numverify API, you’ll need to click the ‘Configure’ button to go straight to ‘Configure’ tab.

Ensure Code Interpreter is ticked and click ‘Create new action’.

Enabling the “Code Interpreter” in a custom GPT model is important for using APIs and working with data. It lets the model run code to make HTTP requests, handle real-time data, and do complex tasks like parsing and filtering. This feature also helps keep data secure and manage authentication when connecting to external services. In short, it adds the flexibility and power needed to create dynamic, efficient, and secure applications.

Step 4: Connecting a Custom GPT to an API

In Authentication settings, select API. The Numverify API uses an API key, so make sure the authentication type selected is ‘API key’, and then click the ‘Save’ button.

Step 5a: Creating a Schema for a Custom GPT

Start by locating the example Weather schema. Once you’ve found it, select the entire schema text and copy it to your clipboard. This schema will serve as a base or reference for your custom GPT schema.

Open a New ChatGPT Tab

Now, open a new browser tab and navigate to ChatGPT. This separate instance will allow you to work on creating your custom GPT without losing any progress or information from your current session.

Paste the Schema into ChatGPT

In the new ChatGPT tab, paste the Weather schema that you copied earlier into the message box. This action puts the example schema right in the conversation where you can reference and modify it.

Explain Your Objective to ChatGPT

After pasting the schema, add a message to ChatGPT explaining what you’re trying to achieve. You could say something like: “I am working on creating a schema for a custom GPT model. I want it to be similar to this Weather schema, but it should function with the Numverify API.”

Refer to Your Numverify Query from Step 2: Recall Step 2 where you ran a Numverify query. You should have the Numverify API URL and the JSON result from that query. Copy both the URL and the JSON result.

Paste the Numverify API URL and JSON Result: Back in the ChatGPT tab, paste the Numverify API URL and the JSON result. Make sure to provide an explanation for each. For the API URL, you might say, “This is the Numverify API URL I will be using,” and for the JSON result, you could explain, “This is an example of the JSON result from the Numverify API.”

Once ChatGPT has generated the adapted schema for you, incorporating the Numverify API’s functionalities:

Review the Generated Schema: Review the schema provided by ChatGPT carefully. Make sure it meets your needs, especially in how it uses the Numverify API. Look for key fields like phone number validation, country code, carrier information, and any other details you need from Numverify.

Step 5b: Numverify Schema Ready to Paste into ChatGPT

Rather than following the steps with 5a and using ChatGPT to generate a schema, you could simply copy the ready-prepared schema below that we generated earlier.

  "openapi": "3.1.0",
  "info": {
    "title": "Phone Number Validation Service",
    "description": "Validates and retrieves information about phone numbers.",
    "version": "v1.0.0"
  "servers": [
      "url": ""
  "paths": {
    "/validate": {
      "get": {
        "description": "Validate and retrieve information about a phone number",
        "operationId": "ValidatePhoneNumber",
        "parameters": [
            "name": "access_key",
            "in": "query",
            "description": "Your API access key",
            "required": true,
            "schema": {
              "type": "string"
            "name": "number",
            "in": "query",
            "description": "The phone number to validate",
            "required": true,
            "schema": {
              "type": "string"
            "name": "country_code",
            "in": "query",
            "description": "Country code for the phone number",
            "required": false,
            "schema": {
              "type": "string"
            "name": "format",
            "in": "query",
            "description": "The desired format of the response",
            "required": false,
            "schema": {
              "type": "string"
        "responses": {
          "200": {
            "description": "Successful response",
            "content": {
              "application/json": {
                "schema": {
                  "$ref": "#/components/schemas/PhoneNumberValidationResponse"
        "deprecated": false
  "components": {
    "schemas": {
      "PhoneNumberValidationResponse": {
        "type": "object",
        "properties": {
          "valid": {
            "type": "boolean"
          "number": {
            "type": "string"
          "local_format": {
            "type": "string"
          "international_format": {
            "type": "string"
          "country_prefix": {
            "type": "string"
          "country_code": {
            "type": "string"
          "country_name": {
            "type": "string"
          "location": {
            "type": "string"
          "carrier": {
            "type": "string"
          "line_type": {
            "type": "string"

Schema for Phone Number Validation

This schema describes an API (Application Programming Interface) using the OpenAPI Specification (OAS) format, version 3.1.0. It’s for a “Phone Number Validation Service” and provides structured details on how the API works, its endpoints, and the data it exchanges. Let’s break it down:

1. openapi: “3.1.0” 

This indicates the version of the OpenAPI Specification that the API conforms to.

2. info

Provides metadata about the API.

title: “Phone Number Validation Service” – The name of the API.

description: Explains what the API does, i.e., validating and retrieving information about phone numbers.

version: “v1.0.0” – The current version of the API.

3. servers

Lists the base URL(s) for the API endpoints.

url: “” – The base URL where the API is accessible.

4. paths

Describes the available paths or endpoints in the API and their operations.

/validate: A specific endpoint path for validating phone numbers.

get: Indicates an HTTP GET method for the endpoint.

description: Explains the purpose of this endpoint.

operationId: A unique identifier for the operation.

parameters: Lists the parameters that can be included in the query string of the GET request.

access_key: A required API key for authentication.

number: The phone number to validate (required).

country_code: An optional parameter for the country code of the phone number.

format: An optional parameter to specify the response format.

responses: Defines the responses for the operation.

200: Indicates a successful response and its content format.

5. components

Provides reusable components for the API.

schemas: Defines the structure of data objects used in the API.

PhoneNumberValidationResponse: Specifies the schema for the response of a successful phone number validation.

It’s an object with properties like “valid” (boolean), “number” (string), “local_format” (string), and others, each describing a piece of information about the validated phone number.

Step 5c: Using the Created Schema

Now that you have the adapted schema provided by ChatGPT, copy it to your clipboard. 

Navigate to the environment where you are configuring your custom GPT and paste it over the example schema. 

configuring your custom GPT and paste it over the example schema

Step 6: Testing the Schema and Previewing the GPT

With the adapted schema now pasted in your working environment, click the ‘Test’ button.

schema test for validate phone number

After clicking ‘Test’ you will notice messages in the Preview section of the screen, in the preview screen click ‘Allow’, to allow ChatGPT to access the Numverify API at APILayer.

preview of allowing Custom ChatGPT App to access the Numverify API at APILayer

You may get a response like the one below where an error occurs because ChatGPT is trying to access the Numverify API without an API key:

error occurs because Custom ChatGPT App is trying to access the Numverify API without an API key

Provide ChatGPT with your API key

Provide ChatGPT with your API key

You should get a better result where ChatGPT returns the verification details for the test phone number in the API URL.

Custom ChatGPT App returns the verification details for the test phone number in the API url

In preview you should be able to test the number validation with other phone numbers now that you have provided the GPT with an API key.

Step 7: Deploying The Custom GPT

Before finalizing and saving your progress, it’s essential to provide a privacy policy URL. This step ensures transparency and compliance with data handling and privacy standards. This is crucial when dealing with user data and external APIs.

You won’t be able to save the GPT without one. This URL should link to the privacy policy associated with the API you are integrating into your GPT model. 

Then find the ‘Save’ button,, you need to decide how you want to deploy your GPT application. You’ll typically be presented with several options:

  • Only Me: Select this option if you want the custom GPT to be accessible only to you. This is a good choice if you’re still testing or if the GPT is for personal use.
  • Anyone with a Link: Choosing this option means that anyone who has the specific link to your custom GPT can access it. It’s a suitable choice for controlled sharing, where you want to give access to specific individuals or groups without making it public.
  • Everyone: If you select this option, your custom GPT will be publicly accessible to anyone. This is ideal if you’re creating a tool or service that you want to be widely available.
publish setting button for Custom ChatGPT App

Congratulations! Your custom GPT is now deployed and ready for use. 

Step 8: Running The Deployed GPT

To run your deployed GPT, select ‘Explore GPTS’ from the main ChatGPT menu. 

ChatGPT menu for creating a Custom ChatGPT App

Click the button ‘My GPTs’ to see a list of your GPTs.

create button for Custom ChatGPT App

Double-click the icon of your GPT from the list to run it.

icon of your GPT from the list for Custom ChatGPT App

Remember you will need to provide your GPT with your Numverify API key to run it.

Custom ChatGPT App with your Numverify API key

Congratulations on completing this detailed journey of developing and launching your unique GPT model. 

Your achievement in creating a custom GPT model is a powerful demonstration of the ability to merge the sophisticated features of conversational AI with the essential services of external data sources, similar to the functionality provided by services such as Numlookup to verify a phone number.

As you progress, keep in mind that the potential is boundless. Starting with Numverify API integration represents just the initial step. You have the opportunity to enhance your custom GPT model by incorporating additional APIs from API providers such as APILayer, addressing a wide array of needs and applications.

Continue to push the boundaries, innovate, and use your custom GPT model as a portal to uncover new horizons in the realms of AI and data integration. Your journey is far from over; it marks the commencement of an exciting exploration in the field of AI applications. Wishing you the best of luck and enjoyable coding experiences!

Custom ChatGPT App: Conclusion

Creating a custom GPT model for phone number validation in ChatGPT is a fun and simple project. You don’t need any coding experience. By following easy steps, you can make an AI that uses the Numverify API to check if phone numbers are valid.

First, log in to your paid Numverify account to use its features. Run a Numverify query to see how the data looks. Then, go to the OpenAI platform to create your custom GPT model. Set up the basic settings and integrate the Numverify API by adding your API key and endpoints.

Your custom GPT can now validate phone numbers and provide detailed information. This makes data management easier and improves customer interactions. You can also add more APIs to expand your GPT’s abilities. This project shows how combining AI with external data can open up many new possibilities. Enjoy creating and exploring your custom GPT!

Custom ChatGPT App: FAQs

What is a custom ChatGPT app?

A custom ChatGPT app is a special version of ChatGPT made for specific tasks. It can use other services, like Numverify, to do things like check if phone numbers are valid.

Do I need coding skills to create a custom ChatGPT app?

No, you don’t need coding skills. You just follow simple steps and use easy tools to make it work.

Why should I add APIs to a custom ChatGPT app?

Adding APIs helps the app get real-time data and do more tasks. For example, the Numverify API lets your app check if phone numbers are real, making things easier and more accurate.

How do I start making a custom ChatGPT app?

To start

  1. Log in to your paid API account, like Numverify, and get your API key.
  2. Go to the OpenAI site and create your custom GPT model.
  3. Enter the API key and settings.
  4. Test your app to make sure it works, then use it!

Sign Up for free at Numverify to get your access key and create a custom ChatGPT for number verification.

Sign up free - Numverify phone number verification API