BitButler
SEA Web3 Concierge. Banking at the speed of voice.
Created on 9th January 2026
•
BitButler
SEA Web3 Concierge. Banking at the speed of voice.
The problem BitButler solves
Why We Built This
Managing your money should feel like speaking to a trusted assistant—not wrestling with software built by engineers for engineers. BitButler exists to make blockchain invisible.
The World We Refuse to Accept
Picture this: A mother in Jakarta wants to send money to her son studying abroad. She opens her banking app, navigates through seven screens, enters an IBAN she copied from WhatsApp, double-checks every digit with mounting anxiety, pays a $25 fee, and waits three business days. She does this because she has no other choice.
Now picture the crypto alternative: That same mother must first understand what a "wallet" is. Then memorize (or dangerously store) 24 random words. Then figure out gas fees. Then pray she didn't accidentally send funds to a contract address. Then wonder why her transaction is "pending" for twenty minutes.
This is the state of finance in 2026. And it's absurd.
3.9 billion people use mobile banking. Only 420 million use crypto. The technology that promised to democratize finance has instead built a higher wall—one made of jargon, complexity, and fear.
What If It Didn't Have to Be This Way?
What if you could manage your entire financial life the same way Tony Stark manages his?
"JARVIS, transfer $500 to the house account."
No endless screens. No confirmation dialogs. No anxiety. Just you, your voice, and an AI that understands not just your words—but your intent.
How BitButler Changes Everything
BitButler is not a wallet. It's not an app. It's not another tool you have to learn.
BitButler is your personal financial AI—a super app that listens, understands, and acts.
Speak naturally:
"Send 50 dollars to Mom."
"What did I spend on food this month?"
"Move half my savings into stablecoins."
BitButler confirms your intent, then a quick biometric scan—Face ID or fingerprint—and it's done. Real security without the friction.
Under the hood, BitButler handles everything else:
- It knows who "Mom" is because you told it once, months ago
- It converts your intent into blockchain transactions with 99%+ accuracy
- It pays gas fees invisibly through smart paymasters—you never see them
- It batches operations to save you money you didn't know you were losing
- It executes on Lisk L2 with sub-second finality, so it feels instant
You don't need to understand Account Abstraction. You don't need to know what ERC-4337 means. You don't need to care about Layer 2 rollups.
You just need to speak—and confirm with a glance.
Who Is This Really For?
The Overwhelmed
You've heard about crypto for years. You know it matters. But every time you try to learn, you hit a wall of complexity that makes you feel stupid. You're not stupid. The interfaces are. BitButler removes the interface entirely.
The Time-Starved
You're a professional with 47 tabs open and three meetings in the next two hours. You don't have time to navigate DeFi dashboards. You need to move money now, and you need it done in five seconds. BitButler: one sentence, one glance, done.
The Underserved
You live in a region where banking infrastructure is unreliable, fees are predatory, and cross-border transfers take a week. You deserve the same financial superpowers as someone in Singapore or Switzerland. BitButler doesn't care where you live.
The Visually Impaired
Most financial apps are accessibility nightmares—tiny text, complex navigation, zero screen-reader optimization. BitButler is voice-first. If you can speak and hear, you can use it.
Your Parents
They call you every time they need to do something on their phone. They're terrified of pressing the wrong button. They've heard about "Bitcoin scams" and want nothing to do with crypto. But they can have a conversation and use their fingerprint. And that's all BitButler needs.
The Vision
In 2011, Siri made voice assistants mainstream. In 2016, Alexa brought them into our homes. In 2022, ChatGPT made AI conversational.
In 2026, BitButler makes blockchain invisible.
WeChat Pay onboarded 900 million users by making payments feel like sending a text message. M-Pesa transformed Kenya by making banking as simple as a phone call.
We're not building another wallet. We're not adding another app to your home screen.
We're building the last financial interface you'll ever need—because the best interface is no interface at all.
The Bottom Line
People don't want crypto. They don't want blockchain. They don't want DeFi.
They want to send money to their kids. They want to save for retirement. They want to split the dinner bill without awkwardness. They want to feel in control of their financial lives.
BitButler gives them exactly that—with nothing more than their voice and a fingerprint.
"Hey BitButler, send a hundred dollars to my daughter."
"Sending $100 to Sarah. Please confirm with Face ID."
One glance. Done. Secure.
Challenges I ran into
Challenges I Ran Into
1. The Passkey-to-Blockchain Translation Problem
The Bug: React Native's Passkey APIs return a P-256 (secp256r1) public key—but Ethereum and Lisk use secp256k1. These are fundamentally different elliptic curves. You cannot just "plug in" a FaceID credential to sign blockchain transactions.
The Journey: We spent two days down a rabbit hole trying to make
react-native-passkey
output secp256k1 keys. It's cryptographically impossible—the Secure Enclave hardware only supports P-256.The Solution: We pivoted to ERC-4337 Account Abstraction with a WebAuthn validator. Instead of forcing the hardware to speak Ethereum's language, we deployed a Smart Account that understands both. The on-chain
ECDSAValidator
verifies P-256 signatures natively. The Secure Enclave stays happy, Ethereum stays happy, and the user never knows any of this is happening.2. Lisk Sepolia's Missing ERC-4337 Infrastructure
The Problem: When we started building, Lisk Sepolia had no deployed ERC-4337 EntryPoint, no Bundler, and no Paymaster. Account Abstraction simply didn't exist on the network.
The Panic: Our entire architecture depended on AA. Without it, we'd have to fall back to seed phrases—the exact thing we were trying to kill.
The Solution: We used Pimlico's public Bundler which supports Lisk Sepolia, and configured
permissionless.js
to route UserOperations through it. For the EntryPoint, we verified that the canonical v0.7 deployment existed at0x0000000071727De22E5E9d8BAf0edAc6f37da032
. Crisis averted.3. The Viem + React Native Polyfill Nightmare
The Error:
ReferenceError: TextEncoder is not defined
The Context: Viem is designed for Node.js and browsers. React Native is neither. It lacks
TextEncoder
,crypto.getRandomValues
, and several other Web APIs that Viem assumes exist.The Fix: We installed polyfills in a precise order at app entry:
import 'react-native-get-random-values'; import 'fast-text-encoding';
These must load before any Viem import. One wrong import order and the entire app crashes on launch. We learned this the hard way—three times.
4. Voice Recognition Latency vs. Transaction Confirmation
The UX Problem: Speech-to-text takes 500ms–2s. Parsing intent takes 200ms. Biometric auth takes 300ms. Blockchain confirmation takes 1–5s. Stack these linearly and you get 7+ seconds of waiting—an eternity in UX terms.
The Insight: Users don't mind waiting if they feel progress. Silence is the enemy, not latency.
The Solution: We designed a multi-stage feedback loop:
- Listening → waveform animation + haptic pulse
- Thinking → "I heard: send 50 to mom" (instant text echo)
- Confirming → Slide-to-confirm + biometric prompt (parallel)
- Executing → Progress ring + transaction hash preview
- Done → Success chime + confetti animation
Total perceived wait time dropped from 7s to ~3s—not because we made it faster, but because we made it feel alive.
5. The "Who Is Mom?" Contact Resolution Problem
The Challenge: When a user says "send 50 to Mom," we need to resolve "Mom" to a wallet address. But crypto has no universal contact book. ENS exists, but adoption is low. Phone contacts exist, but they don't have wallet addresses.
The Workaround: We built a local contact alias system stored in
expo-secure-store
. Users can say "Remember that Mom is 0x123..." once, and BitButler stores the mapping encrypted on-device. For the hackathon demo, we pre-populated "Mom" and "Sarah" as mock contacts. In production, this would sync with an on-chain social graph (Lens, Farcaster) or ENS reverse lookups.6. Expo SDK 54 + React Native 0.81 Compatibility Hell
The Symptom: Random crashes, missing native modules, and build failures after upgrading Expo.
The Cause: Expo SDK 54 shipped with React Native 0.81, which introduced breaking changes to the native module system. Several dependencies (
react-native-passkey
,react-native-reanimated
) hadn't updated yet.The Solution: Pinned specific versions that played nice together, ran
expo prebuild --clean
to regenerate native projects, and prayed to the dependency gods. It works now. We don't touch it.7. The Gasless UX Without a Funded Paymaster
The Ideal: Users speak, biometric confirms, transaction executes. No mention of "gas."
The Reality: Paymasters cost money. On testnet, there's no incentive for anyone to sponsor gas for strangers.
The Compromise: For the hackathon, we pre-funded a minimal sponsorship via Pimlico's testnet faucet. In production, the business model would require either:
- A freemium tier (first 10 gasless tx/month)
- A fee-on-transfer model (0.1% of transaction value)
- Enterprise sponsorship (merchants pay for their customers' gas)
Tracks Applied (1)
Winner Classification
Technologies used