DALL-E’s API is now available in public beta and it’s a great time to ship some cool iOS apps.
Since a Swift implementation is missing from DALL-E’s API docs, I thought of building a project and playing around with URLSession and multipart form-data, Vision face detection, and of course SwiftUI.
The demo application detects face using Vision request, masks it using CoreGraphics and runs the DALL-E API using the masked image + original image + a prompt.
The results are okayish…
Here’s my GitHub repository with the complete source and an accompanying tutorial on Medium.
PS: While I try to figure out my weekly newsletter plans, I’d love to know what y’all have been reading lately?