top of page
Logo_COCA_New (1).png

COCA x Aurora AMA Recap: Self-Custodial Banking in Practice

  • Aug 21
  • 8 min read

On August 19, COCA CEO Vasili Paulau joined Aurora Labs CEO Declan Hannon for the second episode of Connected, Aurora's interview series with the teams it builds with.


They spent an hour on the practical side of self-custodial banking: why private key export was hidden for months and is coming back, what Aurora Intents changed for deposits and $COCA purchases, what happens when a transfer goes to the wrong place, and more.


Missed it? Here are the key takeaways 👇



🔐 Self-custody came first, the banking app came second


Vasili has been in crypto since 2019, and self-custody was the part he fell for immediately.


"I was really in love with the idea that we move the ownership of the funds away from some sort of centralized organization, which can be a bank, exchange, etc., to you as a user."


The idea came before the product. While working with Wirex as a client, he met the people who became his co-founders, and the question they started from was a simple one: why not build something that looks and behaves like the banking app you already use, but where you own the funds, and nobody can block you or cut off access.


Development started in 2025 with a small MVP.



🚪 You get the app before you get the KYC form


Onboarding is the first thing any regulated product runs into, and Vasili called it a constant job rather than a solved problem.


COCA launched in Europe only, at a time when the MiCA rules were still taking shape. Two things shaped the flow from the start.


The first is a product decision. Many neobanks ask you to pass KYC before you see anything at all. COCA does the opposite: you get into the app and can try self-custody first, and verification comes at a later step.


The second is not a decision at all. Which features work in which country is set by local regulators, not by preference.


"It's not only our desire to limit certain functionality or access within certain regions to the product, but it's basically the regulations."



⚖️ Two rulebooks instead of one


Declan, who worked at Revolut before Aurora, put the difference plainly: a traditional fintech has one broad set of regulations to satisfy. COCA has to satisfy fintech rules and crypto rules at the same time.


Regulation also changes how people behave. When USDT ran into trouble in Europe, COCA saw a large shift from USDT to USDC across deposits and withdrawals.


It reached the rewards too. Cashback was paid in USDT from May last year, users pointed out what MiCA meant for them, and COCA switched. You now pick your reward currency: USDC, or EURC for euro payments.


Part of the work is simply explaining. Telling a crypto-native person what self-custody means takes a sentence. Explaining account abstraction to a national authority takes considerably longer.



🔑 Seed phrases: the problem nobody has solved yet


This was the most honest stretch of the conversation.


How do you give people full ownership without handing them something they can lose forever?


"I don't think there is one company on the market right now who completely solved this puzzle. Right now, there is a compromise here between security and simplicity."


When COCA moved to full self-custody in April, the team deliberately hid private key export until it could work out how to communicate the risk, particularly to users who are new to Web3. The feature is coming back to the app now, and the approach is iterative: ship the team's version, collect feedback from users, adjust.


Longer term, Vasili would rather not show a key at all. Inheritance and social recovery, closer to what Argent did in its early days, are being explored as an alternative, possibly next year.



🔄 What Aurora Intents actually does


Aurora Intents sits on top of the NEAR Intents protocol and runs through a solver network. Instead of picking a route yourself, you state the outcome you want, for example USDC in your COCA account paid for with SOL, and solvers compete to fill it. They compete on cost, speed and finality, not on price alone.


Inside COCA it powers two things.


Buying and selling $COCA in the app


The first attempt at this was a DeFi flow built on an Aerodrome pool, and the team decided not to release it.


"You are not getting the amount you want to... as well as the price impact was significant, meaning that you can pay twice the money you're able to pay from other sources."


Getting a good price that way needs deep liquidity, and liquidity takes time to build. Users needed $COCA now.


What shipped instead runs entirely inside the app. COCA shows you how much $COCA you need for the next tier and what that tier gets you, so you do not have to work out the amount yourself. One tap, confirm, and it is paid from your USD account. No external wallet, no manual swap. Selling works the same way, straight back to your USD account.


Cross-chain deposits


The deposit list grew with the integration, and one addition stood out: USDT on Tron, which Vasili named as the most requested network COCA did not previously support, and which matters most to users outside the European Union.


You pick the asset and the network, get an address, and send from a wallet or an exchange.


Withdrawals are the next step. Today you can withdraw crypto in USDC and EURC. Support for Bitcoin, Ethereum and other assets is planned, along with more networks on both sides.



🛡 What happens when a transfer goes wrong


Assets sent to the wrong address are one of the largest sources of loss in crypto, and there is usually nobody to raise a ticket with. Three things in the current setup reduce that risk.


One address across EVM networks. Non-EVM chains such as Tron and Solana have their own, but there is far less to get wrong.


Deposit addresses persist. They can change for technical reasons, and the team still recommends checking yours before a large transfer. If it has changed and you send it to a previous one, the funds are not lost.


Stalled swaps are retried, then refunded. If an intent cannot be filled, the solver network retries it, and if that does not work the funds go back to you.


"There is always a space for an error, right? What we as a company offering the end product, we should take care about these errors."



💸 How tiers work, and where the rewards come from


$COCA is the fuel for the tier system. You buy the token, lock it, and the tier unlocks the perks.


Depending on the tier you get cashback, subscription rebates in categories such as AI tools, music and YouTube Premium, and benefits from partner services. The highest tier, Elite, pays 8% cashback on eligible transactions. Some transaction types are excluded, which is what keeps the programme sustainable. Full conditions are in the tier system T&C.


Nothing is taken from you. Locking is not a payment.


"You get it, you stake it, you get the perks. Don't want the perks anymore, you withdraw, get COCA back."


Leave the tier and your tokens come back to you, ready to hold or sell, and the cashback you earned along the way stays yours.


The part that rarely gets said out loud is where the rewards come from. COCA does not fund them out of its own pocket.


"We are not paying the rewards from our pocket, but we are using the existing DeFi tools."


Balances are converted to stablecoins and earn through protocols such as Morpho. There is no custody to maintain and no branch network to pay for. Vasili contrasted this with the traditional model: take ten dollars from you, lend nine of them out, earn on them, and pay you back a small share of that revenue.



🏦 Where crypto banking is behind, and where it is ahead


Behind, in one word: adoption.


"Our brain is made in a way that we prefer sticking to the current tool, even if it's not favorable in your regard."


Switching a financial product is a genuinely hard decision. The job is to give people a reason to look at an alternative at all.


Ahead: the infrastructure. Self-custody is not only about ownership, it is about what ownership makes possible. Vasili expects that within the next decade users will move between different financial products depending on the use case, carrying a single wallet with them.


There is also what self-custody removes. Branches build trust, but Vasili called that trust partly an illusion, since banks with large offices have failed before. Declan added that a rescued bank still gets paid for, through taxes rather than through lost deposits. A bank run works because the bank is holding your money. It does not work the same way when you are.



🤖 Why the AI Agent Is Still an Internal Demo


COCA already has an internal demo where you chat with an AI and it books a ticket or a table at a restaurant. It works. It is not shipping yet, and the reason is security.


"We should be 100% completely sure that within a certain vulnerability which we are not aware of, the user funds cannot be withdrawn."


Vasili expects self-custodial wallets to get there first, because they sit closest to new technology, followed by DeFi protocols, and only then fintech apps. Intents suit this well: you state an outcome instead of a sequence of manual steps, which is much closer to how you would talk to an agent.


For COCA the plan for the next 12 months is deliberately narrow: AI automation for recurring payments and in-app bookings, introduced step by step, with the user confirming each action, at least in the early versions.



💬 Community Q&A Highlights


1. What is the biggest challenge COCA is solving right now?

Entering new markets. Issuing a card is only part of it. Replacing someone's local banking app means supporting local payment instruments and helping people move funds over from their existing bank, ideally without a separate licence in every country.

Part of the payments infrastructure and the bank partners behind COCA was built by Wirex. The products target different users: Wirex tiers are built around holding a large portfolio inside the app, while COCA tiers are based on the token. Card and banking features are similar in places, and Aurora Intents support is one of the differences COCA has added recently.

Yes. Declan tested this on himself, standing in a queue with two fish he had decided to buy and a card he had not topped up. The deposit landed and was spendable in roughly 10 to 15 seconds.

That is handled at the Aurora Intents layer. A transfer that does not go through is retried, and if it still cannot complete, you are refunded.

Planned, without a date. Balances are converted to stablecoins today so they can earn APY. The direction is a separate ETH account where assets stay in ETH, earn through Morpho and similar protocols, and back an on-chain loan you spend against. You would then repay from other assets, a USDC deposit for example, without touching the ETH.


"So, short answer, yes, we do plan to add this. When? Depends on several factors."

No. Aurora Intents is how funds get into your COCA account, in the same way you would transfer between two bank accounts. The payment itself goes through COCA and the card.

Complex cross-chain movement first, in Vasili's view, along with staking and yield decisions in DeFi. Anything touching payments comes after the security question has a proper answer.



🚀 Looking Ahead


Aurora Intents is live now for deposits and for buying and selling $COCA in the app. Here is what was named as coming next, all of it without dates:


  • More assets and networks for deposits, and withdrawals beyond USDC and EURC.

  • Private key export returning to the app, with inheritance and social recovery being explored as the longer-term answer.

  • An ETH account with on-chain lending on top, so you can spend against ETH instead of converting it.

  • AI automation for recurring payments and bookings, with confirmation at every step.

  • New markets, local payment instruments, and a simpler way to move over from a traditional bank.


Watch the full session for the whole conversation, and follow COCA and Aurora on X.

 
 
 

1 Comment


cellesim
Aug 27

The discussion on the practical side of self-custodial banking is incredibly insightful, especially the challenges of balancing user ownership with ease of use, like the seed phrase dilemma. Ensuring that users have constant, secure access to manage their funds, wherever they are, is crucial for the very concept of self-custody. For individuals looking for seamless ways to stay connected and manage their digital assets, even in places like Mali, https://cellesim.com/en/esim-mali offers essential mobile data options. As an individual connected with Cellesim, I see how important such reliable connectivity is for financial autonomy.

Like

Join over 1M+ crypto users worldwide

ALREADY EARNING CASHBACK, YIELD,
AND REAL-LIFE PERKS WITH COCA

image 96 (1).png
image 96 (2).png
bottom of page