SkiFi

Case studies, Shuriken, Web3 and DeFi

Shuriken: from social alpha to executed trade, in under 500ms

A social alpha to trade execution platform for sophisticated on-chain traders, across Solana, Base, BSC and Monad.

Overview

Shuriken is a social alpha to trade execution platform built for sophisticated on-chain traders. It connects to any Discord server or Telegram group, surfaces token mentions as actionable signals, and lets traders execute across Solana, Base, BSC, and Monad, without leaving the interface.

The platform spans four touchpoints: the Shuriken Terminal (web), a browser extension for one-click trading on any supported site, a Telegram bot for mobile-first position management, and an AI Agent Kit that trades autonomously within user-defined guardrails.

The challenge

Crypto alpha moves in milliseconds. The tools traders were using moved in minutes. The typical power trader was juggling a stack of disconnected tools: Discord and Telegram for alpha, DexScreener for charts, a separate wallet for execution, and a Telegram bot for automation. Every context switch cost critical seconds, and in memecoin trading, seconds are everything.

  • Signals buried in Discord noise, no filtering by caller reputation or market cap
  • Manual copy-paste of contract addresses from chat to execution interface
  • No way to set entry rules in advance; everything was reactive
  • Mobile experience was an afterthought, position management required desktop
  • No path to automation without writing bots or hiring devs

The core insight

The problem was not signal quality. Traders already knew where good alpha lived. The problem was the distance between seeing a call and acting on it.

Signal architecture

Making powerful boolean logic feel like building a filter, not writing a query. Each signal is a named rule with a source (a specific Discord server, a channel, a Telegram group) and a set of conditions. Conditions chain with AND logic by default, OR logic on demand. The UI makes this feel like layering filters, not constructing logic trees.

  • Source selection with live preview of recent messages
  • Condition pills that expand into parameter editors
  • Real-time count of matching signals from the last 24h
  • Named presets: "New gem under $500K", "Top caller alert"
  • Under 500ms from message to feed, an engineered constraint, not aspirational
  • Signal attribution: which rule fired, which caller triggered it

Alpha

Alpha is the main landing experience and Shuriken’s biggest differentiator from every other terminal on the market. It is an infinite canvas where traders build their own real-time signal environment, not a generic feed. Alpha is not just a screen, it is a layer: no matter which page you are on inside the terminal, your full alpha feed is one gesture away.

  • Create multiple columns, each streaming a different source: a Discord server, a Telegram group, a Twitter/X list, or a custom group assembled from specific people across platforms
  • Each column is an independent, focused stream, not a firehose
  • Toggle between Messages view and Tokens view with a single tap
  • Add, remove and reorder columns in real time without losing your stream state

Token analytics

Every token surfaces a consistent data layer: price chart, market cap, liquidity depth, holder distribution, and transaction history. The analytics view was designed to answer the three questions traders ask in the first 10 seconds of evaluating a call: is it liquid, is it distributed, is it moving?

  • Holder concentration heatmap, spot whale risk at a glance
  • Transaction feed with wallet tagging (dev, sniper, whale)
  • Price chart with on-chart indicators and zoom levels
  • Bundled buy detection and deployer wallet history

Trading interface

The terminal collapses the trader’s entire stack into one view: live alpha feed on the left, token analytics in the centre, order panel on the right, and a curated X stream inline. Exit strategies are configured at the point of trade, not after, and enforced without manual monitoring.

Track wallet and Profile

Instead of a flat list of addresses, Shuriken introduces the Profile, a named identity that can hold multiple wallets and link to social profiles across Discord, Telegram, and Twitter/X.

  • Group multiple wallets under a single named Profile, see the full picture of a person, not just an address
  • Attach social identities so the on-chain and social person are unified
  • When someone calls a token, instantly pull up their Profile to see every wallet they control and every call they have made
  • Track patterns over time: who consistently hits early, who pumps and dumps, who calls blind

MemeSlot

MemeSlot is the memecoin scanning layer, a three-column layout built for traders who live in the high-velocity world of meme tokens. The interface is deliberately fast and dense: scroll quickly, evaluate in seconds, act immediately.

Group owners dashboard

Community owners, Discord server admins and Telegram channel operators whose communities generate alpha, get a dedicated analytics layer.

  • Top callers and low performers, ranked by call accuracy, volume generated, and community engagement
  • Token distribution metrics: which tokens your community calls most and what the hit rate is
  • Most trending and most profitable tokens across all your group’s calls in a given period
  • Message activity and engagement analytics

Browser extension

One-click trading on any supported site, without breaking the trader’s workflow. The extension injects a compact trading panel onto DexScreener, DexTools and others, so traders never have to leave the page they are already on. The design challenge was a hard constraint problem: inject a full trading interface without obscuring the chart data the trader needs to make the decision.

  • Buy and sell directly from DexScreener
  • Real-time PnL overlay on positions header
  • Preset amounts and wallet selection per chain
  • TP/SL automation attached at execution

Reflection

Designing for speed means accepting that elegance is a constraint, not a goal.

There is no established playbook for designing systems that turn social noise into trading signals. Every pattern had to be invented from first principles.

For autonomous systems, the UX of safety is as important as the UX of capability. Users will adopt powerful tools if, and only if, control feels real and legible.

Every element in a high-density trading interface either earns its space by reducing friction or costs the user attention they could be spending on execution. The edit made most often was removal.

Outcome

  • $200M+ total trading volume in the first 6 months
  • Under 500ms signal-to-screen latency
  • 4 chains at launch, 60s AI Agent Kit setup
  • Signal delivery was the most-cited differentiator

What we did

  • Product Design, Product Strategy, Prototyping