Description

I wanted to write a tool in Rust which takes an Image as command line argument and then sends this image to ChatGPT to generate an image description. Once it has received the image description it uploads the image to a WordPress media library as new image. After the upload it will add the image description received via ChatGPT as the ALT text and description of the uploaded image in the WordPress media library.

The script will take the URL of the Wordpress site, the username and application password for Wordpress as well as the OpenAPI URL, ChatGPT model and OpenAPI API key as configuration parameters from a .env file.

Prerequisites

To use this application you need an account for ChatGPT and generate an access token so you can use the OpenAI API. Instructions can be found at their documentation site

In addition, you should have a WordPress installation with a recent version of WordPress and an application token configured for your user account. This allows programmatic access to your wordpress instance. This can be found under Users -> Profile -> Application Passwords

preparation

Install Rust on your machine. Copy the .env.example to .env

The .env file is either expected in the current working directory OR in $HOME/.image-helper-wordpress/.env

testing

cargo run -- /path/to/your/image.jpg

compilation

cargo build --release

will build the binary in ./target/release/image-helper-wordpress

running

 $image-helper-wordpress: ./target/release/image-helper-wordpress /Users/username/Pictures/DJI_0247.jpg
Generated description: Aerial view of a circular lake surrounded by dense forest, with a small cluster of houses lining one edge. A second smaller lake is visible in the background, also encircled by forest.
Successfully uploaded image to WordPress with description
Image URL: https://example.wordpress.com/wp-content/uploads/2024/11/DJI_0247-scaled.jpg