A strong everyday card for crypto users, with a Hong Kong company holding your balance

Updated October 2026

Licensed

Available in US

Fast payouts

18+ Only

If you have trouble getting an ordinary bank card — because of where you live, what documents you hold, or simply because no mainstream bank will have you — the Card.fi card is worth putting on a shortlist. It converts crypto into a spendable balance and gives you something you can use at a shop, in a café, in an app and, if you want, at an ATM. My score for it is 6.6 out of 10. That is not a bad score, and it is not a grudging one either. It is the score of a card that is unusually good at the job it does and unusually quiet about the job it does not do: the job of telling you, in plain language, who is holding your money.

The conditions described here are current as of October 28, 2026.

Read on and you will find what the card actually charges, how you load it, where it is weak, and what I would tell a friend in their position. If you are choosing between crypto cards on price alone, the short answer is that Card.fi is not the cheapest to open and not the cheapest to spend, but it is one of the few that tells you your fees before you pay. If you are choosing on the basis of who stands behind your balance, this page will be less comfortable, and I will not pretend otherwise.

Verdict

Card.fi scores 6.6 out of 10. Read that number as two different cards living inside one product. On the mechanical side — can I get one, can I load it, can I spend it — it is excellent, and it earns the top marks available in my scoring. On the question of what stands behind the money, it is thin, and it earns the lowest mark in the table, not because the company looks like a fraudster but because the card’s own public pages do not name the licences that would tell you what protections you have.

The score table below shows every criterion with its mark and the reason behind it.

Two audiences in one sentence. This card suits you if you are spending many different coins on a virtual or metal card, you are comfortable converting crypto into a company-held balance, and you want a card that works in the places ordinary cards do not. It does not suit you if your priority is deposit protection, a named banking partner, or the ability to walk into a branch with a problem. It also does not suit you if you want to hold value on the card rather than spend it — the withdrawal fee and the seven-day hold are not designed for that.

The Cost mark of 2 out of 10 needs unpacking, because on paper a card with no monthly fee looks cheap. The issue is not the monthly fee; it is that opening a virtual card costs $49 on Classic, and the fees you pay on every transaction sit on top of that. A $49 opening fee on a card you might use for a few hundred dollars of spending is a large share of the whole relationship. The Premium card, at $249, halves the top-up fee and trims the currency-conversion fee by a point, which is meaningful — but only if you genuinely spend enough to earn it back.

Will they give you the card?

The short answer is: probably, unless you live somewhere the service is not permitted.

The banner says cards ship to “200+ countries”. The terms set one exclusion: residents of, or orders from, “a country subject to sanctions that would prohibit the service”. That is the only country rule the card states. There is no citizenship-versus-residence distinction beyond it — the card does not say it checks your nationality as against where you live, and it does not publish a list of the countries it considers prohibited.

For a reader who has been turned away by ordinary banks, this matters. A card that ships to 200-plus countries with one stated exclusion is a genuinely wider door than most. But I want to be careful about what that sentence does and does not tell you. “Ships to 200+ countries” describes delivery, not approval. It does not promise that an application from any of those countries will be approved. If your worry is whether your own country’s rules allow you to hold this product, that is a question for a local adviser rather than for a banner.

Sanctions screening is not a formality even when the card treats it as one. Note that under standard anti-money-laundering practice, nationality is recorded and screened separately from residence, because screening and issuer risk policy are often applied to a person’s citizenship and country of birth as well as to the country on their address. If your citizenship is on a list somewhere, you may find a door closed that the “200+ countries” line suggests is open. The card’s public pages do not say how it screens, or what happens if a check fails.

The one thing the card does state plainly is this: the issuer may require identity verification before issuing a card or raising limits. If verification fails, the card is not issued and the fee is refunded. Minimum age is 18. Documents are not stated — the public pages do not list which documents to send, in what form, or how long a check takes. That is a practical gap, and it is worth asking support before you send money.

So: if you are over 18, resident somewhere the service is permitted, and you are willing to identify yourself, the card is open to you on paper. What happens after that is what the next sections cover.

How to get it

Applying is straightforward in principle: you create a Card.fi account, choose a card, verify your identity if asked, and wait. The interesting part is which card you choose, because the card does not offer one option but three families, and the differences between them are more than cosmetic.

The available cards are Virtual Classic and Virtual Premium; physical Classic and Premium; and metal in stainless steel. Every physical order includes a virtual card, which is useful if you want to start spending before the plastic arrives. Cards are issued “under licence from the relevant card schemes” — but the card does not name those schemes. It does not say Visa, it does not say Mastercard, and it does not state what currency the card carries. Those two omissions are small on their own and quite large together, because they determine both how widely the card is accepted and what happens to you when a shop’s payment system refuses to recognise it.

For delivery: physical Classic and Premium ship within 5–7 business days; metal is shipped express. Replacement cost is not stated. The shipping footprint is the same “200+ countries” as the banner.

Who gets a card at all is set by the identity verification noted above, and I found no tiered verification levels described on the public pages — that is to say, the KYC level required for a Virtual Classic and the level required for a metal card is not stated. If the amount matters to you, that is a question worth asking in writing before you pay $49 or $249.

Here is the practical shape of the choice, in the order you are likely to make it:

• If you want to try the product, Virtual Classic at $49 is the cheap door in.
• If you spend heavily or travel, Virtual Premium at $249 lowers both the top-up fee and the currency-conversion fee, and comes with higher limits.
• If you want something in your hand, any physical order ships in 5–7 business days and includes a virtual card from day one.
• If you want metal, you get a dedicated account manager and a larger ATM allowance, and you need to ask what it costs, because the price is not listed.

Prices for physical and metal cards are not stated. That is not a small gap: it means the two most tangible versions of the card are the two you cannot price from the public pages.

Fees: what it really costs

The card’s fee structure is the part it does best at explaining. There is no monthly fee, no annual fee, no inactivity fee and no minimum-balance fee. That sounds like the beginning of a good story. It is not, because the real costs sit in three other places: the one-time issuance price, the fee on every top-up, and the fee on every currency conversion.

The table below sets out every fee and limit the card’s public pages state. Values I could not find are marked as not stated, which means exactly that — the pages do not say.

Now the part that matters, which is what this table does to an ordinary purchase. Take a coffee costing $5. You will have topped the card up first, so the top-up fee already applied — 1.5% plus a dollar on Classic, which on a $50 top-up is $1.75, and on a $500 top-up is $8.50. Then, if your balance is in one currency and the coffee is priced in another, the 3.5% FX fee on Classic takes $0.18 off that $5. On Premium the same coffee costs $1.20 less at the top-up step and $0.05 less at the conversion step. None of these amounts is dangerous on its own. What matters is that they are separate charges, they stack, and the card does not fold them into a single number for you.

That said, I want to put the Cost mark in proportion. The card is not expensive to keep. It is expensive to open, expensive to fill, and moderately expensive to use across currencies. A card you open once and use daily for local spending in a single currency will feel almost free by the third month. A card you open and use twice, in two currencies, will feel expensive, because the $49 never amortises.

What you get for the Premium tier is exactly this: higher limits, lower conversion fees, and priority support. Metal adds a larger ATM allowance, a dedicated account manager, and higher cashback. The tiers are laid out like this:

One number in that last table deserves a separate word. Cashback is paid on card spending only. Classic pays 0.8%, capped at 10 USDT a month. Premium pays 1%. Metal pays 1.2% with caps of 100 USDT a month and 5,000 USDT, or 2% with caps of 250 USDT a month and 20,000 USDT. Separately, a headline rate of 10% at coffee shops is capped at 10 USDT a month. And on balances you can earn “up to 18%” on Metal — variable, which in this context means it goes down as easily as up. Treat the 18% as a possibility rather than a plan, and note that the headline 10% at coffee shops is the kind of figure that gets quoted out of context by people selling the card.

How to Top up

This is where the card earns its highest marks, and where it is genuinely hard to criticise.

Card.fi supports 32 coins across 5 networks: Ethereum, Base, BNB Chain, Solana, and Kaanch. Among the coins listed are USDT, USDC, ETH, BNB, SOL and KAANCH. Each card has its own deposit address. The amount is converted at the market rate on arrival. The top-up fee is the one in the table above — 1.5% plus $1.00 on Classic, 1% on Premium — and the terms show a conversion fee before each top-up, which is unusual and welcome.

Four things about this setup deserve your attention, in ascending order of how much they matter.

The first is simply breadth. Five networks and 32 coins means that for most people, whatever they already hold can be sent straight across without a swap, an exchange account, or a bridge. If you already sit on USDT on Solana, you are one transfer away from spending.

The second is the per-card deposit address. This is good practice. It means a support agent can see which transfer belongs to which card, which is exactly what you want when something goes wrong.

The third is the network mismatch risk, and this is the one that can cost you everything. Sending funds on a network the provider does not support, or to an address belonging to another network, is a well-known way to lose a deposit outright. Blockchain transfers are irreversible; the network cannot reverse them and cannot freeze them. The card’s own material flags this — sending on the wrong chain risks permanent loss. One trap worth naming specifically: if you hold USDC and assume the popular cheap version will do, check carefully, because Circle stopped supporting USDC on Tron in February 2024, so “USDC over TRC20” is not a native option.

The fourth is what the card cannot tell you. The minimum top-up is not stated on the public pages — it is shown in the app or account. The order completes after on-chain confirmation, so there is a wait between sending and being able to spend, and a slow network means a longer wait.

Taking money back off the card has its own costs and its own clock. Withdrawals are paid in a supported stablecoin 1:1, with a fee of 1% of the amount, and there is a seven-day hold during which the withdrawal can be cancelled. Closing the card returns funds to a held balance that is still withdrawable. Virtual cards cannot be cancelled once issued — a small line with a large implication, which I will return to in the risks section.

Paying with it every day

This is the strongest part of the card, and it is why I gave Everyday use the full marks.

There are virtual, physical and steel cards. Apple Pay and Google Pay are supported on Classic and Premium. Physical cards support ATM withdrawals, with Premium and Metal carrying larger allowances than Classic. Cards are managed in the Card.fi account, a web dashboard. Spending limits are described as “standard” on Classic and “higher” on Premium, with no figures given, and you can set limits yourself in card settings. Freezing a card from the account takes effect immediately.

Apple Pay and Google Pay matter more than they sound. When you add a card to a phone wallet, the card number is replaced by a device-specific number stored in the phone’s secure hardware, never on the company’s servers, and each payment carries a one-time code. In plain terms: you can tap to pay without handing over a card that could be copied, and you can kill the card in your account without hunting for the physical thing in a drawer.

Now the constraint you will hit soonest, and it is not the fees: blocked categories. The card blocks 18 merchant category codes. A merchant category code is simply the four-digit number a business is classified under when it starts accepting cards, so blocking a category is the card declining that whole class of merchant by rule. The blocked list includes:

• crypto exchanges (6051) and quasi-cash (6050)
• cash and ATMs (6010, 6011)
• stored-value loads (6540) and remittances (4829)
• financial institutions (6012) and securities brokers (6211)
• betting and gambling (7995; also 7800, 7801, 7802, 9406)
• telemarketing (5966), dating (7273), massage (7297)
• political and religious organisations (8651, 8661)

Two of these deserve a word. Blocking 6051 means you cannot buy crypto with this card — which is logically consistent, since you fund it with crypto in the first place. Blocking 6011 means the ATM category is restricted on cards where withdrawals are not included, which is another way of saying virtual cards do not do cash. And the gambling blocks are a policy choice rather than a legal requirement: many issuers block 7995 by policy, and the blocked list varies by issuer, so a card that blocks gambling is not unusual. But if gambling or a specific merchant class matters to you, this list tells you in advance rather than at the till.

One more everyday-use fact worth having: a declined authorisation can leave a hold on your funds for several days, in some cases as long as seven. That is standard for card payments, but on a prepaid balance it feels sharper, because the balance is smaller and a hold eats a larger share of it.

And what happens when something goes wrong is not covered here. Card.fi does not address transaction disputes in its terms; disputes fall under the issuer’s cardholder terms, and complaints on card accounts are passed to the On5 Card Network. That is a real answer — it tells you where to look — but it means you are reading someone else’s document, and the protection you have is the protection that document gives.

What can go wrong

Four risks, in the order I would worry about them.

The first is who holds your money. This is the one that gave the Reliability and complaints mark its 5 out of 10. The card is the trading name of Kaanch Limited, Hong Kong, under Hong Kong law. The cards are issued by the On5 Card Network. On-chain settlement is by the Kaanch Network. Licences are not named — the card does not tell you which licence it holds, from whom, or for what. And the balance you hold is custodial: the terms describe it as “a sum we owe you”. Safeguarding and segregation of customer funds are not addressed in the terms.

Take that seriously for a moment. If your balance is a debt of the company, then your money is not a customer deposit with a deposit guarantee scheme. It is a claim against Kaanch Limited. In the EU, an electronic money institution must safeguard customer funds in dedicated accounts or insure them, and even then those funds are not bank deposits and may not be covered by a deposit guarantee scheme — so even a well-regulated version of this product gives you less protection than a bank account. Here, the licence that would tell you whether any of that applies is not named. That is the gap, and no amount of good press release closes it.

The second risk is history. Custodial platforms have failed before, and not in small ways: Mt. Gox reported about US$350 million stolen across 2011–2014, Coincheck about US$400 million in 2018, and the Ronin Network US$625 million in 2022. The pattern is not always fraud; it is often that a single company held too much and could not move it. A company’s licence existing does not make this risk zero.

The third risk is that the card is a programme, not a network. The card is issued under licence from the relevant card schemes, but Visa and Mastercard themselves do not issue crypto cards to consumers — they license issuers, and they can suspend a programme. That has happened: Visa terminated its global agreements with FTX, announced 13 November 2022, and the FTX US debit card was wound down by its issuer. Visa also suspended WaveCrest in Europe in 2018, closing all WaveCrest-issued Visa programmes. The shape of a closure is usually a notice period, then loading and spending stop, then you are asked to withdraw what is left. There are also unconfirmed reports of a 2026 closure in this category where new loads ended 8 July 2026 and spending stopped 8 August 2026 — a single secondary source, so I am not treating it as fact, but it shows the pattern is current.

Now the fourth risk, and it is the one that has no fix. Blockchain transfers are irreversible. A wrong-network deposit may be lost outright, or need manual recovery with no guarantee of success. Stablecoins are not censorship-proof either: Tether says it disables tokens in wallets on the OFAC sanctions list and has frozen funds tied to illicit activity. And USDC is not a guarantee of anything — stablecoins can and have lost their peg, as TerraUSD did in May 2022 and USDC did briefly in March 2023 during the US banking crisis. The card’s value proposition assumes stablecoins hold their value between the moment you deposit and the moment you spend. Usually they do.

What I could not find is worth stating too. I found no official incident involving Card.fi itself. Searches turned up no Trustpilot page and no Reddit discussion for Card.fi or Kaanch — the Trustpilot check was run on 2026-10-21 and returned nothing. That absence cuts both ways. It may mean a young, clean product; it may also mean very few people use it. Store ratings were not found either. And there is a structural asymmetry worth naming: if something goes badly wrong here, you will be dealing with a company in Hong Kong under Hong Kong law, through email and Telegram, with no branch to walk into. The card promises an acknowledgement within two working days and a resolution or explanation within fifteen, with priority for Premium. Fifteen days is a long time when your balance is frozen.

The final sharp edge: virtual cards cannot be cancelled once issued. Combined with a custodial balance and a seven-day withdrawal hold, that is a structure in which getting your money back takes a deliberate process rather than a click.

What to take instead

I want to be careful here. None of the alternatives below is clearly better; they score lower on my own scale, and I would not tell anyone to switch on the strength of a comparison table. They are worth knowing about because they show the same category has different shapes.

Cardex scores 5.6 out of 10. It offers one crypto card for “200+ countries” with no country list or exclusions stated, and runs on the VERVE, Visa, Mastercard and UnionPay networks with virtual and physical cards. Its App Store listing is Naira-focused, and conversion happens at the moment of payment: the site describes an “AI-Powered Conversion Rate” that scans global exchange rates and picks the best one as you spend. No fee percentage is stated on the site text I read on 2026-10-22, which is exactly the sort of thing to ask about before you commit. Funding is by USDT, USDC and BTC wallets, networks not stated.

CardFlow scores 4.9 out of 10. It is aimed at Nigeria — “Nigeria’s smartest fintech card platform” — and at “Nigerians and global users”, with no country list or exclusions stated. The card is a virtual USD Mastercard created in about 3 minutes, and creating one is described as completely free in its site FAQ. Fees are not stated, and funding is by Bitcoin, USDT, ETH or a Naira bank transfer, networks not stated.

HyperCard scores 4.7 out of 10. Official pages state no country list; the App Store page says the card works in “more than 176 countries” — which is a statement about acceptance, not about who is eligible to apply, a distinction worth holding on to across all of this category. It is a prepaid card with a balance in USD or EUR, funded by depositing crypto such as USDT, BTC or ETH in the app. Networks and fees are not stated on the pages I could read.

Placed side by side, the honest comparison is this:

The pattern is unflattering to the category and flattering to Card.fi: for most of these, the fee is the thing you cannot find out. Card.fi publishes its numbers. It also tells you its card currency is not stated, its physical and metal prices are not listed, its ATM allowance amounts are not stated, and its dispute terms live in someone else’s document — but the fee schedule is there, in full, before you pay. For a reader whose main problem is uncertainty rather than a few tenths of a percent, that is worth more than a cheaper card with an unpublished fee.

There is also a category of alternative that is not on this table because it is not really comparable: a self-custody-linked card, where the funds stay in your own wallet until the moment of payment. The trade is symmetrical — you keep control, and you take on sole responsibility for key loss, phishing, and clipboard-hijacking malware that silently swaps a copied address. Whether that trade suits you depends on how you store things, not on which card you buy. A seed phrase, typically 12 to 24 words, is an unencrypted form of the private key; anyone who obtains it can restore the wallet and spend the funds, and there is no built-in recovery because the network cannot freeze or return them.

Who I would recommend it to

I would recommend this card to a specific person, and I can describe them. You hold several different coins, mostly stablecoins, across networks like Ethereum, Base, BNB Chain or Solana. You want to spend them without first converting everything to a single currency at an exchange, and you want the thing to work in a café, in an app, and sometimes in a currency that is not your own. You are over 18, resident somewhere the service is permitted, and you are comfortable identifying yourself to a Hong Kong company that will hold your balance as a claim against it. You will top up in meaningful amounts, not in five-dollar dribbles, because that is how the fee structure makes sense. And you will set spending limits yourself, because the card’s own limits are described only as “standard” or “higher”.

I would not recommend it to a person whose first question is “what happens if this company goes out of business”. Not because the answer is bad — because the answer is not published, and the licence that would tell you is not named. I would not recommend it to someone who needs to dispute a transaction, because Card.fi’s own terms do not cover disputes and send you to the issuer’s cardholder terms. And I would not recommend it to anyone planning to travel on the balance as a cash reserve: virtual cards do not include ATM withdrawals, the seven-day hold and 1% fee make getting money back slower than you would want in an emergency, and a card cannot overdraw only because the balance stops.

Before you apply, I would do three things. Ask support, in writing, which card scheme the card runs on and in which currency it holds its balance — the two answers that shape every transaction you will ever make. Read the withdrawal terms closely, because that is where the custodial structure shows itself. And put a small amount in first, then a real amount, and check how long the on-chain confirmation takes and how long the funds take to come back out. If those two numbers are acceptable to you, the rest of this card is unusually well built for everyday use.

Frequently asked questions

What is Card.fi?

Card.fi is a crypto payment card. You load it with cryptocurrency, it converts the amount at the market rate on arrival, and the card spends a balance from there. It is the trading name of Kaanch Limited, a Hong Kong company, and the cards are issued by the On5 Card Network. It is not a bank, and your balance is described in its terms as a sum the company owes you rather than a bank deposit.

How does Card.fi card work?

You create an account, order a card, verify your identity if the issuer asks, then send crypto to the deposit address assigned to that card. The 32 supported coins across 5 networks are converted at the market rate when they arrive, and the order completes after on-chain confirmation. From then on the card behaves like a prepaid card: it spends only what you have loaded, so it cannot overdraw.

Is Card.fi a bank?

No. It is a card programme run by Kaanch Limited in Hong Kong, with cards issued by the On5 Card Network and on-chain settlement by the Kaanch Network. Your balance is custodial, meaning the company holds it, and the public pages do not name the licences behind the programme. That is the single most important fact to weigh before you fund it.

Who owns Card.fi?

The card is presented under the trading name Kaanch Limited, registered in Hong Kong and operating under Hong Kong law. Licence details are not named on the public pages, and I could not find them. If who stands behind the company matters to you, that is the question to ask support in writing.

Does Card.fi require KYC?

The issuer may require identity verification before issuing a card or raising limits. If verification fails, the card is not issued and the fee is refunded. The minimum age is 18. Which documents are needed, and how long a check takes, is not stated on the public pages — so ask before you pay.

How much does a Card.fi card cost?

A Virtual Classic card costs $49 one time and a Virtual Premium card costs $249. There is no monthly, annual, inactivity or minimum-balance fee. Beyond that you pay the top-up fee — 1.5% plus $1.00 on Classic, 1% on Premium — and a currency-conversion fee of 3.5% on Classic, 2.5% on Premium. Physical and metal prices are not listed.

Is Card.fi card free?

Not at the start. The cheapest route in is a Virtual Classic at $49. CardFlow, by comparison, describes creating a virtual card as completely free, so if the opening cost is your main obstacle, that is worth a look. What you save on Card.fi you partly give back in per-top-up and conversion fees.

How do I top up Card.fi card?

Send one of 32 supported coins on one of 5 networks — Ethereum, Base, BNB Chain, Solana or Kaanch — to the deposit address assigned to your card, then wait for on-chain confirmation. Coins listed include USDT, USDC, ETH, BNB, SOL and KAANCH. The critical warning: sending on the wrong network can lose the funds permanently, so check the network twice before you confirm.

How do I withdraw money from Card.fi?

Withdrawals are paid in a supported stablecoin at 1:1, with a fee of 1% of the amount, and there is a seven-day hold during which the withdrawal can be cancelled. Closing the card returns your funds to a held balance that remains withdrawable. Virtual cards cannot be cancelled once issued, so think of the card as one-way in that respect.

Is Card.fi card safe, and can I use it in my country?

On safety, honestly: your balance is held custodially by a Hong Kong company, the licences are not named, and the terms do not address safeguarding or segregation of funds. That is a real risk, not a formality. On country, the banner says “200+ countries” and the terms exclude sanctioned countries where the service is prohibited — but shipping reach is not the same as approval, so check locally before you apply.

Content created by the «cryptocardcheck» team