# Launching My First API Subscription: Fashion Classification Made Easy
Written on
Chapter 1: Introduction to API Development
If you share my passion for machine learning, you might have pondered the possibility of monetizing your AI model. Throughout my academic journey, I've developed numerous models, yet I never considered the idea of transforming one into an API and publishing it on RapidAPI, a marketplace where users can subscribe to utilize your service. Recently, I decided to test this concept with my latest AI model and gauge its market viability.
Section 1.1: The AI Model in Focus
Recently, I embarked on a machine learning project aimed at classifying fashion items into their correct categories. This model was built entirely from the ground up, utilizing a unique dataset I compiled, distinct from commonly used datasets.
My dataset is notably superior, featuring images of human models in clothing rather than just the garments themselves. This approach enabled me to create a more realistic and generalized model, achieving an impressive accuracy rate of 85%. The image below demonstrates the model's performance.
Image by author
For those interested in the dataset details, I invite you to check out the following blog post:
I Created a Custom Fashion Dataset for Deep Learning: 52,000 labeled fashion images from 10 categories.
Section 1.2: Developing the API
Upon completing the Fashion Classification model, I promptly developed an API and listed it on RapidAPI. I established four subscription tiers: a free plan that permits 150 API requests at no cost, along with three paid plans featuring varying request limits.
Fashion Classification API Documentation | RapidAPI
This API enables users to predict clothing item categories by uploading images.
rapidapi.com
Chapter 2: Marketing the API and Initial Success
After launching the API, my next step was to promote it and see if there would be any interest in the paid subscriptions. I shared my API in various Reddit communities, highlighting its advantages, especially the unique dataset used. Within the first few days, I recorded 50 users subscribing to the free plan, and just four days later, I secured my first recurring paid subscription!
The first video discusses how to build and sell your own API, providing insights into the process and potential earnings.
The second video outlines four simple steps to create and sell an API, making it accessible for everyone.
In response to my initial question, I can confidently affirm that there is indeed a market willing to pay for thoughtfully designed APIs. Looking ahead, I aim to enhance the API by expanding the dataset with more images from diverse clothing categories and refining the AI model.
If you're interested in receiving lifetime access to this API, please leave a comment below—I would be thrilled to gift it to the first ten individuals. I hope this post has inspired you on your own journey!