-->Unigox Blog | What we learned building a neobank from scratch

&larr; Back to blog<br>Published on 25th of May 2026 by Unigox Team<br>A year and a half ago we showed up at a hackathon in Chicago with a stack of ideas about crypto and AI, and walked out convinced we were going to fix peer-to-peer trading with zero-knowledge proofs.<br>Twelve months later we're shipping NeoBank-as-a-Service. This is the story of what happened in between, and what we wish someone had told us before we started.<br>TL;DR<br>Building a neobank yourself takes roughly eighteen months and $300K+ in compliance and engineering before your first transaction settles.<br>The hard part isn't writing code, it's stitching together twenty+ providers, four kinds of compliance coverage, and an App Store review cycle.<br>Modern fintechs win by automating the back office, not by having more rails than anyone else.<br>If you have an audience already waiting, time-to-market beats margin every time.<br>We built NeoBank-as-a-Service for the audience-first case: branded app on the App Store in about four weeks, on the same rails Unigox runs on.<br>The first idea: kill fraud in P2P with ZK-TLS<br>The pitch was beautiful. Anyone who's traded on a P2P exchange knows the painful part: the seller releases crypto from escrow only after manually confirming the fiat arrived. That gap is where scams happen, where disputes pile up, and where merchants get burned.<br>ZK-TLS (zero-knowledge transport layer security) lets you prove cryptographically that a fiat payment landed in a bank account, without exposing the bank login or the account contents. Plug that into a P2P flow and the manual "yes I got the money" step disappears. Trustless, instant, automated. We were going to put peer-to-peer trading on rails.<br>Three things killed it.<br>ZK-TLS only works against banks with a web interface. No web banking, no proof. We started in developing markets, and the users there overwhelmingly bank from their phones. Nobody trusts the browser. We could have shipped to the US where everyone has Zelle and Cash App, but those are exactly the markets where P2P doesn't matter.<br>Every bank integration was bespoke. Nigeria alone has thousands of microfinance banks. Each one needs its own ZK-TLS harness. We did the math: the scaling curve is sand.<br>The market we were saving was already shrinking. While we were heads-down building, central banks across our target geographies were quietly legalizing crypto on/off-ramps. The big exchanges followed, and now we ourselves have a corporate bank account in Nigeria to buy and sell crypto directly. The thing that made P2P necessary, the lack of any other path, was going away.<br>There was a poetic twist. We built a system to protect P2P users from getting scammed, and then spent months getting pen-tested by people trying to scam us. Lost real money trying to help people not lose money. Pink glasses, as we called them later.<br>What we learned about the actual market<br>Once we stopped defending the original idea and looked around, the bigger market was sitting right next to us: cross-border payments and branded neobanks running on stablecoin rails.<br>The pattern we kept seeing:<br>A founder with real distribution. A community, a following, a niche audience that already trusts them.<br>An obvious money-movement use case for that audience. Remittances. Tuition payments. Commodity settlement. Local-to-local FX.<br>Zero appetite to spend eighteen months wiring up compliance, KYC, AML, ramps, custody, App Store reviews before serving a single user.<br>These people don't want to build payment infrastructure. They want to launch a branded app and start onboarding the audience they already have.<br>The hidden anatomy of "just integrating an API"<br>Here's the part that surprised us the most, and the part that almost nobody talks about publicly.<br>When you say "I'm going to build a neobank," you imagine: pick a payment API, plug it in, ship an app. In reality you're looking at a checklist that runs something like:<br>A legal opinion for App Store submission<br>A compliance officer and a written AML program<br>A regulatory license, or partnership with someone who holds one<br>A KYC vendor for user onboarding<br>A KYT vendor for ongoing blockchain transaction monitoring<br>One or more fiat on/off-ramp partners, often different ones per corridor<br>An IBAN-issuing partner for receiving and sending fiat<br>A custody model (custodial or non-custodial, each with different obligations)<br>Liquidity prefunding across every chain and every fiat account you operate<br>A rebalancing system so you don't run out of inventory where users want to withdraw<br>A back office that can review flagged transactions in hours, not days<br>And then the actual app, with payments UX, error states, and customer support<br>Before you process your first transaction, you're already six figures into legal, compliance, and integration work. The 18-month figure isn't dramatic. It's what it took us.<br>The provider gauntlet<br>The compliance side is hard. The vendor side is worse.<br>Every payment...