Don’t be afraid of AI; try to adapt. The extraordinary development of artificial intelligence creates concerns about job loss for many people. This is true, but what’s even more true is that AI cannot directly take away your job. People who use AI at a good level can stay one step ahead of you and take your job. Therefore, we must improve ourselves in this field. With OpenAI’s release of the plugin usage for ChatGPT, a new level has been opened. We should quickly adapt to this. I will explain the stages of developing a plugin for ChatGPT.
ChatGPT plugins consist of 3 things:
- An HTTP server
- An OpenAPI specification
- A manifest file
Steps to Create a Plugin in OpenAI
- First, create an API server using ChatGPT
- Tell ChatGPT about your requirements
- Also specify the library in which the server will be created (in this case, FastAPI)
- In this step, create the plugin manifest.
A plugin is a json file that includes:
- Simple metadata about the plugin
- How the plugin will be presented to a human
- Also, how it will be identified to the language model
ChatGPT plugin manifests.json
{
"schema_version": "v1",
"name_for_human": "TODO Manager"
"name_for_model": "todo_manager",
"description_for_human": "Manages your TODOs!",
"description_for_model": "An app for managing a user's TODOS"
"api": { "url": "/openapi. json" },
"auth": { "type": "none" },
"Logo_Url": "https: //example.com/logo.png",
"legal_info_url": "http://example.com"
"contact_email": "hello@example.com"
}3 . Then, deploy the server and notify it of the json file.
4. After distributing the plugin server, add the plugin to ChatGPT
- Provide the domain where the plugin is hosted
- Provide authentication token if necessary
5. Finally, install and test the plugin as a user.
We compiled this resource from the Twitter posts of dear Rohan @clusteredbytes, we thank Rohan
Leave a Reply