I turned my TypeScript UI system into a motion pipeline for @StackingDao.
No After Effects. No 3D tools.
Just pure SVG, their brand kit, and my jhosephe skill.
1/ Pressing the coin triggers a debossed green S ignition and 0-100% ring charge.
Citi cuts its Bitcoin target to $99,532. Wall Street measures market friction, while open protocols measure computational truth. Price targets are temporary opinions; decentralized architecture is permanent infrastructure.
58
Jhosephe retweeted
One stBTC bought 1 BTC in September. Today it buys 1.00447.
Bitcoin Staking rewards on @Stacks compound into stBTC every day.
Jhosephe retweeted
Engagement is the most misleading number in crypto.
A coin can sit at the top of every trending list and still have almost nobody talking about it. One account with a big following posts, a few thousand people quote it, replies pile up and the chart looks like a movement, it’s a megaphone.
Then there’s the opposite, no single post goes huge, but thousands of different accounts are independently posting about the same thing, quieter, easier to miss and a lot closer to what a real narrative looks like.
Same total attention, completely different thing.
The part of lunarcrush i think people sleep on is that it tracks both sides. Engagements (how much reaction is happening) and creators (how many distinct accounts are posting and actually getting engagement), across @X, @Reddit, @YouTube and @tiktok_us, put them side by side and you can finally ask “is this loud, or is this wide?”
To be clear, that doesn’t tell you where price goes, nobody should pretend it does. What it tells you is what kind of conversation you’re looking at before you decide how much to trust it.
What’s new is you don’t have to dig through dashboards to ask, lunarcrush has an official connector in @claudeai, backed by an mcp server that serves the same data as plain markdown. So you can type “pull engagements and active creators for these two topics over 30 days, which one is a megaphone?” and it fetches live numbers instead of guessing from training data.
Also, it isn’t just crypto same question works on a stock, a sports team, an artist’s album week, 35+ categories. The tool doesn’t care what the conversation is about, only what shape it has.
Most of us read the timeline like it’s the market, it’s really a sample of the loudest accounts, this is a way to check how big the actual crowd is.
Most market terminals look identical: dark neon noise trying way too hard to feel “real-time.”So I reimagined @LunarCrush as archival paper instruments & anodized telemetry hardware.100M+ posts/day. 41 categories. 6 networks — all rendered in 60fps vector motion.A breakdown thread 🧵👇
5
3
333
"Read by machines. Made by people."Zero After Effects keyframes were touched here — every piece is pure code-driven SVG motion built with the Jhosephe UI design & motion system.The full design + SVG-to-60fps-MP4 agent skill is now open and downloadable here for builders & designers: jhosephe.vercel.app/skillsHo… the @LunarCrush team digs the tribute 🚀
2
35
5/ Last one pairs FinancialChart + AudioMeter into a 28x10 green LED matrix tracking @StackingDao analytics: $78.21M TVL, 641.98 BTC, and $4.09M in rewards paid.
Rendered via TypeScript + Playwright + FFmpeg at 1620x1080, 60fps.
Which style fits DeFi best?
1
13












