Exploring OpenAI: My Journey from Article Writing Tools to Business Chat Integration
Sebastian Ardila Agudelo
Software Engineer experienced in React, TypeScript, Node.js, Figma and Low-code Development
Introduction to OpenAI API and Its Opportunities
I have been learning about the OpenAI API to build personalized chatbots for businesses, which comes with a lot of opportunities that people should take advantage of. While learning about it, I delved into embedding vector search, which has given me a lot of understanding about semantic search. I'm building a chat to integrate many businesses, and I'll provide you with constant updates about how I'm creating these features. Keep reading to know more about it.
Creating Tools for Article Writing
First of all, in my job search, I found a copywriting job that pays me for articles. However, I am a software developer, so as you might guess, I started by creating some tools to improve the time and quality of writing those articles.
For this, I created some automatic prompts, passing variables to them and executing those prompts directly from my interface to the OpenAI API using Node and React.
I know it was not the best UI, but i didn't need to do that, it was just a personal tool. As you can see, some texts are in red for professional purposes, but the idea was that I could create all the texts I wanted and regenerate those texts each time I wanted without writing the prompts again. Every 'Generate' button was connected with the GPT models. This tool saves me 1 hour in creating the article with minimal costs.
Developing a Personal Bot with ChatGPT Plus
I have already created a bot of myself in ChatGPT Plus. It was easy; I just had to attach some files to the assistant, and boom! I was able to talk with a professional profile of myself. If you have the Plus version of ChatGPT you can see my bot here: https://chat.openai.com/g/g-faGTBT7sT-sebastian-ardila
领英推荐
Updates on the Beta API for Assistant Bots
Then, in the last few weeks, OpenAI brought us an update about the beta API for assistant bots, which allows developers to have their own assistants by connecting to the API: https://platform.openai.com/docs/api-reference/assistants.
So, I could make my personal assistant available to the public because I had already done that with another assistant I created for one of my favorite coffee shops in the city.
Building a Chat System for Business Integration
I've hidden some options for personal purposes. In the image example, I asked the chat system that I created for the menu of the coffee shop, so I was able to interact with the menu whenever I wanted. Then, I could offer a small solution to people who have a business like that, and that is a personal project I am currently working on.
Technical details
I'll write on Medium in the coming days about how I connected an assistant using Server-Sent Events to replicate what ChatGPT does, delivering information in chunks as if the bot were writing. My next steps involve using the LLMS models to save data and convert that information into vector-embedded data for quick searches.
If you want to know more about it, please subscribe to my Medium profile for more technical details about this information and how I code it. Thanks!
My Medium profile: https://medium.com/@sebastian-ardila
Please feel free to put your comments about this post in the comments.
Fullstack Software Engineer Specialist
7 个月Nice article, unfortunately I don't have the ChatGPT Plus subscription to interact with your bot. Looking forward to read the next articles and progress you do in your project.