A place for developers to talk about building on Farcaster.

/dev

Spent the FarCon week talking with Farcaster devs. A recurring theme: the downsides of deploying Farcaster apps on Vercel using Next.js SSR. Both cost surprises and technical challenges came up repeatedly. Takeaway: convenience isn't free.

Farville 🧑‍🌾

clanker

Mini Apps dev @dtech | Building /hazanlegacy for fun

Chatted with the creator of Eggs @farcasteradmin.eth. His advice? Stay away from SSR. You'd be hitting scaling pain early and SSR added cost and complexity with very little upside. Unless you truly need SSR, you're probably better off without it.

Here’s the kicker: with thousands of daily users, Vercel bills can easily cross into the thousands of dollars per month, and that’s just for the frontend infra. No DB, no analytics. SSR scales your costs linearly. You pay for every request. It adds up fast.

Hazan Legacy

Already refactored /hazanlegacy from Next.js to a Hono backend with a Vite + React frontend. Now it's faster, cheaper, and way easier to host flexibly. Static frontend + API backend is a pattern that just works, especially in Farcaster’s Miniapps.

Funny how these cycles work. We started with backends generating HTML, then moved to SPAs, then onto SSR with Next.js. And now? We're swinging back to decoupled frontends and lightweight backends. The pendulum keeps moving.

Talked to the creators of /farville, /clanker etc. All confided their Vercel bills were shockingly high. Why? They leaned heavily on server-side rendering with Next.js, and usage costs added up fast. Takeaway: heavy SSR will burn money.