monad has a lot going on.
NFTs, meme coins, dApps, campaigns, beginner guides, $MON price - all scattered across X, different sites, different discords. if you're new, you don't know where to start. if you're active, you're still missing half of what's happening.
so i built MonadHub 👉monad-hub.vercel.app
not because someone asked me to. because i needed it myself
what it is
a live dashboard for the monad ecosystem. one site, always updated, easy to scan.
tabs for:
- trending X posts from monad
- hot NFT collections
- meme coins (live data)
- active campaigns
- new dApps and launches
- beginner resources
- $MON price + monad TVL
the goal wasn't just to list links. it was to make monad explorable for someone brand new and for someone who's been here since testnet
how i built it
Next.js + Vercel for the frontend. fast, dashboard-style, built to feel like a real product.
data from:
- twitterapi for discovering monad X posts and campaigns
- supabase for storing curated campaigns with manual approval
- [nad.fun](//nad.fun) API: meme token data
- opensea API: NFT collections
- defillama: TVL + new launches
- coingecko-style APIs: $MON price
campaigns are the most hands-on part:
twitterapi fetches monad campaign posts every few days → saved to supabase as candidates → i review manually → approved ones go live, auto-sorted into hot / ending soon / ended
i didn't want a noisy feed. i wanted something curated enough to trust.
where i got stuck
real talk:
- opensea API calls got too high: had to add aggressive caching before it became a problem
- supabase RLS, enum dropdowns, public read policies: all confusing at first. connecting supabase MCP to codex through PowerShell was its own adventure
- logos and profile pictures kept breaking from wrong image URLs or sizing issues
- font encoding showed weird broken characters for a bit
- analytics couldn't show traffic from before GA was installed, so early numbers had to be estimated from vercel edge logs
none of these are glamorous problems. but fixing all of them is what turned a rough prototype into something people actually use.
AI was my co-builder
i used Claude Code and Codex throughout the entire build.
2 ai tools i trust
not just for generating code, also for debugging visually, connecting APIs, adding supabase logic, improving caching, and pushing changes.
the workflow was: describe what you want or you feel wrong → AI implement/ inspects the code → updates the implementation → test → push.
it's a different way of building. less* *"i know exactly what to write" and more "i know what i want, let's figure it out together"
vibecoding is real. and for solo builders it's actually the move.
what happened after launch
i didn't expect much honestly
then projects started reaching out. builders started dropping into the site to share what they shipped. people said it was the clearest overview of the monad ecosystem they'd found.
close to 500 visitors in early days. more coming in every week.
that felt good. not because of the number but because it meant the thing actually worked. people found it useful. that's the only metric that matters at this stage
what's next?
i have more tools in the pipeline. things i'm building because i need them, same reason MonadHub exists. when you build from a real need, the product shows it.
somewhere in the process of shipping this, i realized i actually like doing this. not just the finished product. the iteration, the debugging at 2am, the moment something finally works the way you pictured it.
i'll keep building. and i'll keep sharing what i learn: the real stuff, not just the hype.
because that's the kind of content worth making.
---
if you're building on monad and want your project listed, reach out.
and if you're a creator or builder thinking about shipping something - just start. the tool doesn't have to be perfect. it has to be useful.
thanks for reading 🫶🏻