By the Variant team

At Variant, we’ve been spending a lot of time in the Farcaster ecosystem. We often find ourselves discussing the ideas that we’d like to see – both as users and as investors.

Below is a list of requests for startups that we’ve compiled across the team.

We’re also hosting a Farcaster developer meetup + brunch on Saturday, April 6 at our office in NYC, which you can apply to here: https://lu.ma/p43szrum. These ideas are a subset of what we hope to see getting built, but if something on this list is something you’re excited to work on (or already are working on!), we hope you’ll reach out.

Clients:

In New Opportunities in the Farcaster Stack, Alana wrote about how several open opportunities are to build clients that treat new data types — specifically, frames and tokens — as first-party citizens.

We haven’t seen anything that’s quite cracked either of those yet, but we’d still very much love to. So, it felt worthwhile to elaborate on what those opportunities might look like.

A frames-first client

The goal of a frames-first client would be twofold: enable persistent frames and improve discovery of new frames. 

Today, it’s difficult to revisit a frame – they’re displayed in-feed and treated much the same as posts or videos are. But this doesn’t make much sense: post-related content is consumed quickly and forgotten even quicker.

Persistent frames need a display more like a Home Screen than a feed. There is also likely value in a tagging and search system for frames. This could help users rediscover frames they saw in a feed elsewhere, filter by all variations of frames created using the same underlying tool, compare frames within a given category (e.g. prediction markets), and more.

What might the discovery of new frames look like? Again, viewing frames more like a library of content than ephemeral in-feed posts creates a prompt to look beyond existing social displays. Maybe the right form factor is something akin to Netflix’s “top 10 trending.” Maybe it’s something else. My sense is that there's a lot of inspiration to be drawn from existing display factors, especially from apps that folks might not think of as traditional “social” interfaces.   

- Alana

Token-centric clients

The other new data type is tokens. There are a lot of interesting ways to create asset-centric clients that lead to more delightful consumer experiences.

An example of a token-centric client is one that’s built around a popular community currency. Currently, there are numerous community currencies trending on Farcaster, each having particular rituals around them, like tipping and rewarding content aligned with community values. As such, those tokens have latent curation value and could be utilized in a client to prioritize the feed and content recommendations. (See also: https://warpcast.com/li/0x3f93ef98)

A great benefit is that tokens are programmable. As such, they can be extended and integrated throughout a client in novel ways. One idea might be to leverage Solana’s token extensions. For instance, a developer could experiment with the token reallocation extensions to reward users for “hot streaks” (e.g. if a user posts a certain number of days in a row) or lack of activity. 

But arguably most powerfully, the asset can be used to bootstrap distribution. Getting people to download a new app feels by far the hardest part of building a successful new Farcaster client. The opportunity to make money is a powerful motivating factor. It’s also one that we’ve seen work well in the web2 app playbook – UberEats offers in-app credit, Cash App provides a $5 sign-up bonus, etc. This, combined with the ability to piggyback off of existing token-oriented communities, can help a new client get its first few thousands users.

- Alana & Li

A specific persona-focused client (e.g. degens)

Another approach to building a new client is designing it for an audience that has more specific product needs. For example, there's room to build a client that serves crypto degens as a first-class audience. A client focused on degens might have features that put assets and profit front and center, with trading features, views of trending assets in Farcaster, and aggregation of earnings opportunities. There could be other such personas that a client targets, where the feature set is sufficiently differentiated vs. Warpcast.

- Li

Frames:

Multi-hit frames (games, etc.)

In Multi-Hit Wonders, Li outlined that a new strategy for consumer application builders is to launch multiple successive applications, and use tokens or another mechanism to stitch together community attention across apps. Frames could be suited to this strategy, with multiple mini apps and games launching and capturing attention transiently. 

One of these applications could be a “I’m feeling lucky” frame, which takes some set of randomized content – could be a random Google search, a random Wikipedia page, even a random set of casts on the network – and displays one at random whenever someone clicks “I’m feeling lucky.” This idea could also utilize in-frame transactions, in that every so often someone wins a community token reward and can claim it. 

- Li & Alana

Memecoin creation frame

Memecoins are having a moment in crypto. We think the best memecoins start community-first, rather than asset-first (examples include $MOG, etc.). A frame that makes it easier to create a memecoin from within the social feed can make sense in enabling assets to be issued and traded directly in-feed.

- Li

Trading bots frames

Over the past few months, Telegram bots have facilitated billions in trading volume, with leading Telegram bots like BonkBot reporting over 250k lifetime users and weekly volumes in the hundreds of millions. The demand for Telegram bots showcases that users are actively looking for more convenient and social experiences for trading. There’s an opportunity to build new Telegram bots native to the Farcaster discovery experience that offer customized trading experiences. 

Channel marketplaces

Farcaster channels present an opportunity to bootstrap new types of intra- and cross-channel marketplaces. Imagine if subreddits enabled users to transact inside of their communities, instead of sending users off-platform. Channel social graphs coupled with in-frame transactions could enable  channel-specific marketplaces. Farcaster could also support aggregation of those channel marketplaces, or a headless marketplace layer.

- Mason

Other

Decentralized content moderation and labeling networks

Recently, Bluesky, the decentralized social media app, announced that it was open-sourcing its moderation tool, called Ozone, to let anyone run their own content labeling service. Labels can be informational, topical, curational, or moderational, and users can tune how the app handles various labels.

As Farcaster grows, we think there’s an opportunity to build a decentralized content moderation network where users can contribute labeling data to casts – potentially coordinated and incentivized by a token. Farcaster clients can ingest this data and customize what content to expose, or let users choose, based on those community-driven labels. 

- Li

Discovery/recommendation algorithms

One of the main concerns with centralized social networks is that centralized tech companies control what content users see on their feeds through their own proprietary algorithms. With a decentralized social network like Farcaster, there is an opportunity for third parties to create customized recommendation algorithms or discovery curation, decoupled from the client. 

What this could look like is developers creating models that users can subscribe to or switch for better discovery or reduction of topics such as trading, politics, or memes to put users in control of what types of content they digest. Additionally, other participants could boost relevant posts similar to how traditional ads work on traditional social networks. 

- Li & Jack