We should expand the listicle tool to allow for users to enable the creation of image covers.
We would likely want to:
- Add a checkbox in the content generation portion to enable the cover image generation.
- (OPTIONAL / INVESTIGATE) Prompt ChatGPT to generate a title + subtitle.
- Create a separate field in the
JSON prompt so the top-level object has a cover field with relevant data. Here we would just reference the image name (and any relevant information) to use for it.
- (OPTIONAL / INVESTIGATE) Add two text fields in the image processing step under a
Cover group: Title Subtitle, where a user could input the text they want generated there.
The two ? steps represent that one should be chosen as we see fit.
The first ? is nice to completely let AI decide everything and make it all automatic, but it would cost more due to prompt tokens. The second ? would mean more intervention from user, but is cheaper.
Things to note:
- How do we want to prompt for a cover image generation?
- Do we just want to use the first image that was generated, a random image from the generated list, or create a new image?
We should expand the listicle tool to allow for users to enable the creation of image covers.
We would likely want to:
JSONprompt so the top-level object has acoverfield with relevant data. Here we would just reference the image name (and any relevant information) to use for it.Covergroup:TitleSubtitle, where a user could input the text they want generated there.The two
?steps represent that one should be chosen as we see fit.The first
?is nice to completely let AI decide everything and make it all automatic, but it would cost more due to prompt tokens. The second?would mean more intervention from user, but is cheaper.Things to note: