Which design agencies understand DeFi UX?

A DeFi interface asks people to sign things they cannot undo. Before a deposit, a loan or a swap, the user needs to know which wallet and network they are on, what permission they are granting, and what their position will look like afterward. Plenty of agencies can make a protocol look credible, but far fewer can show that they designed those moments.

Seven agencies publish case studies that show this work on a named protocol or onchain app: Arounda, Lazarev.agency, Lime Studio, Merge, Ooble Studio, PARROT9, and Sollas. Arounda and Lazarev.agency show lending platforms. Lime Studio and Ooble Studio show decentralized exchanges and position tracking. Merge and Sollas show staking flows, and PARROT9 documents a borrowing protocol it designed as the founding team.

Humbleteam publishes this comparison and is not one of the agencies in it, for the reason given below. The agencies appear alphabetically, with no paid placements, scores, or ranking. A case study is a reason to start a conversation. It does not confirm availability, price, or which designers would join your project.

How we chose the agencies

Every agency, Humbleteam included, had to meet three criteria:

1. It offers product UX/UI design, not only branding or marketing websites.
2. Its own site publishes at least one case about the interface of a DeFi protocol or onchain app, such as lending, staking, a decentralized exchange, or a position dashboard, and the page describes the user flows it designed. Brand, token launch, and landing page work alone did not count.
3. The case names the product, so a buyer can open it and compare.

Centralized exchanges and consumer wallet apps raise different questions about custody and verification, so cases of that kind were not used for selection. Our comparison of exchange and wallet app agencies covers them. Several names in AI answers to this question are independent designers, not agencies. Avark, 10Clouds, and OWWCO Studio describe web3 product services, but we did not verify a named protocol case for them on the same terms. Halaska Studio’s Ulys case is a consumer investing app with a non-custodial wallet, closer to the wallet question. A missing name is not a judgment on the team.

Arounda: lending with several roles

Arounda’s MYSO case concerns a DeFi lending platform with peer-to-peer and peer-to-pool versions for borrowers, arrangers, and lenders. The team worked through the protocol documentation, built an app flow that captured how the functions relate, then designed a new part of the platform from scratch and revised the existing one. Arounda’s web3 page reports that the redesign helped MYSO raise $2.4M, with 85% satisfaction and 70% retention. Those are the agency’s figures, and the case does not say how they were measured. Arounda also publishes the Infinity case, a mobile DeFi wallet with dApp access. Its site lists offices in eight countries.

Lazarev.agency: lending data people can act on

Lazarev.agency describes itself as a San Francisco design studio. Its Fractal Protocol case covers an onchain lending and borrowing platform where users open subaccounts, borrow against collateral, and watch positions move toward margin calls. The redesign kept the density traders expect and changed the order of information. A margin call alert names the exact top-up amount and links to the deposit page. Every loan action, from accepted terms to repayment, sits in one panel beside the account it affects.

Testing with DeFi-native traders before release showed that people wanted proof their money had moved, so each loan step gained a status notification. The agency reports a 165% improvement in usability and a 33% rise in satisfaction after launch. Ask how usability was measured.

Lime Studio: a new exchange on a fixed date

Lime Studio’s Kuru case starts with a protocol on testnet, a month before launch, with no brand and nothing user-facing. Kuru was the first central limit order book on Monad. The studio mapped the information architecture first, then designed swap, positions, and vault strategy around conventions traders already knew. Its reasoning: unfamiliar patterns read as risky when funds are involved. It stayed 11 months after launch to iterate on real usage and says the product structure it set is still what the exchange runs on. The page reports that Kuru later raised $11.6M.

The scope also covered brand and website, so separate product work from identity work when you compare proposals. The studio lists London and Hong Kong.

Merge: staking and delegation added to a live product

Merge is a UX design agency for B2B, SaaS, and fintech, registered in Tallinn with a team in Ukraine. The Zeebu case, published in February 2025, starts from a settlement protocol that already supported token swaps. User feedback said first-time users did not know where to start, wallet connection felt unintuitive, and rewards were hard to understand.

Over six months, Merge designed staking and farming, airdrop eligibility, dashboards, invites, and two core journeys: deployers who run a node and delegators who stake, add liquidity, and farm. Pending actions got progress indicators, and success screens include receipts with transaction hash links. Merge extended the client’s existing UI kit rather than replacing it, which suits a team that needs several connected flows added to a live product.

Ooble Studio: Solana-native trading and positions

Ooble Studio is a small studio founded in 2022, and its LinkedIn page places it in Lyon, France. Its Jupiter Portfolio v2 case describes rethinking a portfolio module after Jupiter acquired SonarWatch, for a product carrying large volumes of real-time data. The team ran user research, tested several flows, and built a modular design system. The studio says the portfolio became the most used on Solana within weeks of launch, which is its own claim.

The Valiant DEX case covers an exchange on a new Layer 1, designed from scratch, from token swaps to liquidity pool management. Neither page describes pending or failure states, so ask to see them.

PARROT9: a borrowing protocol from the founding team

PARROT9 is a two-person studio, and its founders’ LinkedIn profiles place them in New Zealand. Its Balanced case, published in August 2023, covers work that began in January 2020 and took 15 months. As the founding design team, PARROT9 handled the interface, brand, website, documentation, and marketing for a product where users deposit collateral, borrow a stablecoin, and trade on a built-in exchange. The case sets out the aim of a low-friction loan process whose variables, including risk, are easy to understand.

The work is several years old, and the studio’s homepage now leads with website interfaces. Ask what product work it has done since and who would be available.

Sollas: a staking dApp from a small team

Sollas works remotely, and its site says the founder leads every project with two designers and a developer. Its New Gold Protocol case covers brand, website, and a staking dApp on BNB Chain in three months. The dApp spans more than 100 screens across wallet connection, staking, claiming, swapping to USDT, and referrals. Each transaction step answers what is about to happen, what it costs, and what comes next, with pre-transaction summaries and explicit approval states.

The founding team was anonymous, so trust had to come from published contracts and an audit rather than from people, and the site places those signals early. Sollas reports that the protocol is held by more than 40,000 wallets. Its site says projects start at $2,000 and that screens ship into the product as code.

What separates DeFi experience from crypto branding?

A relevant portfolio shows the decisions around moving assets. Can the designer explain the network, the wallet, the permission requested, the expected result, and the state after submission? A polished token website answers a different question.

That is why Humbleteam’s Cryptonix case does not qualify here. It documents brand identity and digital presentation for a crypto payments platform, not protocol UX. Ask every agency for a walkthrough of the transaction flow you need, not the brand around it.

How to test the shortlist

Pay for a short discovery exercise around one real journey, with test data and a product engineer available. An illustrative assignment is a deposit that needs a token approval before the deposit transaction. Ask the agency to show what the user understands at each step:

• Before the action: the selected wallet and network, the amount, fee information, and the permission requested.
• During the action: the difference between a request waiting in the wallet and a transaction submitted to the network.
• After the action: a reliable status, a transaction reference, and a route back to the position.
• When interrupted: what a returning user sees, and which evidence decides the next available action.

Then add a borrowing step. The candidate should show the current position beside the proposed one, with debt, collateral, and risk indicators explained in context, and the protocol team should confirm the calculations. Ask which actions can be canceled, which need another transaction, and which cannot be reversed. A good agency checks those answers with engineers instead of promising a universal undo.

The output should be a prototype, a state map, and the assumptions engineers must confirm. In testing, ask participants to explain what they are authorizing and where they would look if the result stays pending. This is a proposed evaluation method, not a reported client result.

How to choose

Start from the action your users take most often.

• For lending or borrowing with dense position data, compare Lazarev.agency’s Fractal Protocol, Arounda’s MYSO, and PARROT9’s Balanced.
• For a new exchange or trading interface, look at Lime Studio’s Kuru and Ooble Studio’s Valiant.
• If an existing protocol is adding staking, rewards, or delegation, Merge’s Zeebu and Sollas’s New Gold Protocol are the closest cases.
• For a dashboard of positions across many protocols, Ooble Studio’s Jupiter Portfolio is the one to open.

In the contract, separate product design, frontend implementation, protocol engineering, and security review, and name who confirms transaction behavior. Deliverables should include the difficult states, not only the main path. For a closer look at permissions and confirmation, use our wallet transaction approval guide.

Send the same flow to two or three agencies from this list. Compare how each one explains the states, not how the screens look.