project screenshot 1
project screenshot 2
project screenshot 3


Spread a little joy and whimsy, make sure you slip slop slap while you're casting by automatically slapping on some noggles to your Farcaster PFP


Created At

ETHGlobal Sydney

Project Description

This project uses Frog, axum and a face detection algorithm to automatically find (hopefully) every eye position of every face in a user's profile picture, and slaps a random set of noggles onto everyone in the picture, providing download links as well as easy to share social compose intents to add a bit of fun to the weekend and your social profiles :)

How it's Made

Using a couple of rust libraries for image processing called rust-faces and axum for the server and image responses and compositing, for frame UI generation and a sprinkle of frog UI to make it not look like absolute garbage. I didn't really have anything I would consider buildable on my own for the weekend lined up as well as some outside commitments so didn't want to be carried too hard in a group, brainstorming late Friday night knowing I'd have maybe 8-12 hours I figured the nouns partner would at least give me an open-ended enough theme and enough of a constraint to be "complete" without feeling like too much. There were some fun problems I spent way too long on, I love rust s o much it's unreal. Fun to be able to test the pre-trained model on whether it could detect some of the whackier, esoteric PFPs on Farcaster. Turns out humanoid images worked pretty well, punks and pixel art NFTs not so much, although @v and @dwr's pfp's which I didn't expect to work too well actually ended up looking pretty good in testing which was surprising!

background image mobile

Join the mailing list

Get the latest news and updates