Skip to content

01Injective Labs2026

True Current: Decentralized Trading MVP

Trading without the orderbook

Rebuilt a decentralized perpetuals trading product end-to-end in one month, from audit to shipped MVP.

Role
Product Designer (sole end-to-end)
Timeline
4 weeks
Team
Sole designer + engineering team
Constraints
One-month MVP, existing architecture, Nuxt UI base, complex wallet/network logic, changing brand direction
Links
Check in Figma ↗ · Live ↗
  • 4weeks to shipped MVP
  • 90%rejected design recovered
  • 3-steponboarding (was fragmented)
TrueCurrent Exchange: live trading screen with chart and RFQ trade ticket

live trading screen

01

Overview

TrueCurrent is a non-custodial perpetuals exchange built on Injective. It replaces the traditional orderbook with a lightning-fast RFQ model: a trader sets their parameters, quotes are gathered from professional liquidity providers, and the trade executes at the best available price, settled onchain in under a second, with zero trading fees, and with funds that never leave the trader's own wallet. Markets span crypto perpetuals, tokenized stocks, FX, commodities, and prediction markets.

I lead product design for TrueCurrent end to end: research, flows, UI, prototyping, the design system, and handoff into an AI-assisted build pipeline.

02

The problem

Opening a first position on a typical DEX means bridging funds yourself on a third-party site, acquiring gas tokens, decoding opaque wallet prompts, and then facing an interface designed for market-makers. Professionals tolerate it because the orderbook is their instrument. Everyone else leaves before their first trade. On an exchange, that isn't a UX complaint, it's the business problem: activation is volume.

TrueCurrent's architecture removed the structural blockers: no orderbook, no gas, no custody. My job was to make those advantages legible, because an invisible advantage converts nobody.

03

Goals

A first-time trader should get from the landing page to a filled position without needing to understand bridges, gas, or orderbooks. An experienced trader should never feel the product was dumbed down for someone else. And every unhappy path (delayed funds, rejected quotes, failed transactions) should build trust rather than burn it.

04

Decision 1: The price is the interface

The orderbook is intimidating, but it performs one honest service: it shows you the market. Remove it, and users immediately ask whether the price is fair.

The trade ticket became the answer. One firm price instead of a wall of depth. And slippage, DeFi's scariest hidden variable, converted into an explicit control: the trader names the worst price they'll accept, and execution either beats it or doesn't happen. The interface's job is to make that guarantee impossible to miss, and to make a rejected quote feel like protection working, not the product failing.

TrueCurrent Exchange: annotated trade ticket with the worst-price control visible
trade ticket, worst-price control visible

05

Decision 2: The bridge you never see

Bridging is where crypto UX goes to die: real money in flight, zero feedback. We pulled deposits inside the product: pick the chain your funds are on, and the journey is handled. Then we designed the waiting as carefully as the happy path: progress a person can actually read, honest states for delays, and clear recovery when something fails. Funding shouldn't require knowing what a bridge is.

TrueCurrent Exchange: portfolio (background for deposit modal overlay)
TrueCurrent Exchange: deposit modal opened over the portfolio
deposit, inside the product

06

Decision 3: Earn the signature

Every wallet prompt is a withdrawal from the user's trust account. So the product shows value before it asks for anything: browse live markets and real prices first, connect when you're ready, and meet each signing moment explained in human language before the wallet ever appears. Connect → deposit → trade, compressed and readable.

TrueCurrent Exchange: old onboarding funnel versus the new one
onboarding funnel, before and after

07

Process: flows before pixels

Every journey above started as a flow map and wireframes; visual design entered only once the structure survived scrutiny. Below, the progression from early flows to shipped UI, the unglamorous middle where most of the real decisions happened.

TrueCurrent Exchange: progression from flow maps and wireframes to the final UI
flow maps and wireframes to shipped UI
TrueCurrent Exchange: flow map of the funding journeys: connect wallet, bridge via CCTP, deposit from an exchange, or fiat on-ramp
funding journey flow map

08

The design system

TrueCurrent's identity is dark-native: an amber accent, with mint and red reserved strictly for long/short semantics, set in Geist, and every decision expressed as tokens in a fully componentized system. Early explorations pointed the identity in one direction; the shipped system went 180° the other way, and because everything ran on tokens, that reversal cost days rather than weeks. The system lives as a structured spec that both engineers and AI agents consume, which keeps shipped UI true to design.

TrueCurrent Exchange: design system sheet with tokens, long/short color semantics and type
tokens, semantics, type

09

Outcomes

TrueCurrent is live on web and iOS: a zero-fee, sub-second, five-asset-class exchange where the first-trade journey no longer requires understanding the machinery underneath. Predictions shipped into real-world event traffic on day one of the World Cup. I walk through adoption and funnel metrics in detail in conversation.

  • 4weeks to shipped MVP
  • 3-steponboarding with bridge

Impact

  • Faster, more responsive interface across the product
  • A recognisable identity the product can be remembered by
  • 3-step onboarding with the bridge integrated into deposit

Metrics

  • Clear wallet, network and transaction feedback throughout
  • 90%-complete design rejection recovered without moving the deadline

Learnings

  • A reusable system the team still builds on today
  • Sequencing and scope: separating execution quality from strategic fit protected the timeline

Synthesis

What it taught me

  1. 01

    the safe path has to be the fast path, or no one takes it

  2. 02

    colour is a signal, not decoration. it belongs on data, never on chrome

  3. 03

    show consequence (liq price, fees, slippage) before the click

Next case studyXapo Wealth