Powerfeed

Demo

About

The problem it solves Explosive growth of Farcaster audience led to numerous challenges to Warpcast and the protocol in general: a) influx of bots posting spammy and low-effort casts b) engagement farming casts dominating the feed c) lack of genuine non-crypto content d) low visibility for quality content creators To tackle these problems, we've come up with an idea of */powerfeed* – social game that supports quality content , based on three building principles: **Make it fun:** We're holding */powerfeed* games weekly encouraging people to cast on a specific topic. Best casts are hand-picked by our mod team and make it to the Main Section of the channel. **Make it easy:** Every user got 3 bolts (⚡️) daily to send to the cool casts in */powerfeed*. Each bolt has its points based on user's power score (= number of power users engaged with user's casts for the last 7d). When a bolt is sent, points are split 50/50 between sender and cast author. **Iterate:** Throughout the Buildathon we were experimenting with different themes, including IRL/non-crypto content (e.g. 'Offchain Summer') and integrations with other successful Farcaster projects ('Build in Public' in a collab with Talent Protocol). Atm we're finalizing the next iteration of the game that would support quality content creation across all channels. **Interim results:** a) 23k users folowing the channel with 8k users actively engaged in games held b) */powerfeed* sits in top-10 most popular Farcaster channels (measured by total root casts and reaction-to-cast ratio; source: https://dune.com/queries/3342391/5600543) c) Many non-power users posting quality content got extra views&engagement as we bumped their casts onto the Main section of the channel

How it's made

Challenges we ran into Throughout the development we faced numerous challenges related to the development of Frames in Farcaster. --- Complicated Frame designs resulted in URLs being too long for Warpcast. For the development of the frames we used Frog framework and one of the solutions for making it possible to render the frames is to use an ImageHandler. However, after trying out the ImageHandler, we detected that there was an issue in the way Frog processes UI components in the ImageHandler (https://github.com/wevm/frog/issues/353). Due to that, the main solution was to keep everything as compact as possible and to substitute complicated UI design with less components and more images if possible, as it keeps the URL short and frame of a lesser size. --- Avoiding time-outs when processing a lot of data in Frames. Another Farcaster-specific challenge is related to the time-out error which can occur if the frame takes longer than 5 seconds to load. This is a common error among frames which process and/or display a significant amount of data. In order to avoid this error, we made sure to remove most of the time-consuming processing from the frame side and process everything on the server-side and store all the relevant data in the database, which can then be retrieved from the frame side without any significant loading times. --- Constantly updating frame data without running into issues with old data being cached. One of the most reliable solutions which we discovered is using a unique link for each user to display the frames with user-specific data. In cases where the data has been changed for the user, a new unique hash is generated. Hence, we avoid any possible collisions with old data by using a unique hash for each user.