Phoenix Wallet Review (2026): Fees, Custody, US Status

For eleven months, the United States government did to Phoenix what no hacker ever has. It took the wallet away from American Bitcoiners.

ACINQ pulled Phoenix from US app stores on May 3, 2024, a few days after the feds indicted the Samourai Wallet developers for the crime of publishing software. The message to everyone building self-custody tools was not subtle: write code that lets Americans hold their own keys and you might get classified as a money transmitter, or end up a defendant. Wallet of Satoshi fled the US market the same year.

Phoenix came back on April 8, 2025, when the regulatory weather changed.

So here is the honest accounting of what you actually get, what it costs, and where it will bite you.

Five things to know before you download it:

• It's actually yours. Phoenix runs a real Lightning node on your phone behind a 12-word seed. ACINQ never holds your keys or your coins. That is rarer in a mobile Lightning wallet than it should be.
• The fee is the tradeoff. You pay a flat 0.4% to send over Lightning, and receiving can trigger an on-chain fee when the wallet has to make room. That is the price of never thinking about channels.
• It's back in the US and running the latest version. The 2024 exile is history.
• It's easy until it isn't. Onboarding is smooth. The day a channel force-closes, the wallet suddenly expects you to understand seeds and sweeps.
• It's the easiest real Lightning wallet, not the cheapest. Want zero fees? Run your own node. Want zero responsibility? Use a custodian and give up your keys. Phoenix sits, on purpose, in the middle.

At a glance: Phoenix Wallet (2026)

What Phoenix actually is

Start with the part that matters: Phoenix is yours.

Most apps that call themselves Lightning wallets are custodians in a trench coat. Wallet of Satoshi holds your coins and hands you an IOU. Phoenix does not. In ACINQ's own words it is "a real, self-contained Lightning node that runs on your phone," your funds sit behind a 12-word seed only you control, and any standard wallet can recover them. Not your keys, not your coins is the whole game, and Phoenix is one of the few mobile Lightning wallets that actually plays it.

What makes it feel effortless is that it hides the hard part. Every user gets a single Lightning channel against ACINQ's node, and the wallet resizes that channel in place using splicing instead of churning through opens and closes. On-chain and Lightning show up as one balance. You send, you receive, and the plumbing stays out of sight.

That convenience runs on a relationship with ACINQ's node, and to their credit ACINQ says the quiet part out loud. Phoenix, they tell you, is "trust-minimized, but not trustless." Swaps are trustless and an on-chain deposit needs no trust; a channel created from an incoming payment asks for a little trust until it confirms. Either way your money stays spendable, and if ACINQ vanished tomorrow your funds would still be safe. You just could not make new payments until you swept them on-chain with your seed.

The one line every honest reviewer should quote sits in the privacy FAQ: "The current version of Phoenix offers no advantage regarding privacy over existing, hosted, custodial wallets. We (ACINQ) know the final destination and amount of payments." Respect them for writing it down. Also believe them. If chain privacy is your priority, point Phoenix at your own Electrum server or use a different tool.

Is Phoenix available in the US?

Yes. As of right now, an American can download Phoenix and run it. This needs its own section only because so much of the internet is still wrong about it.

The removal was real. On May 3, 2024, ACINQ killed US access, citing recent signals from US authorities that "cast a doubt on whether self-custodial wallet providers, Lightning service providers, or even Lightning nodes could be considered Money Services Businesses." Translated: holding your own keys might make a software company a regulated money transmitter. It landed days after the Samourai indictment, and the whole industry read it as a warning shot.

The return was just as real. On April 8, 2025, Phoenix went back on the US App Store and Google Play, the regulatory posture having softened through 2025. Today the US listing shows the current version, a 4.2-star rating, and no region lock.

One caveat, kept honest: ACINQ never published a detailed legal explanation of the comeback, so treat the reason as a change in the regulatory climate rather than some documented restructuring. The fact that you can install it today is verified.

How Phoenix's fees actually work

Fees are the thing everyone argues about, and the thing most worth getting right. Here is the current schedule, exactly as ACINQ publishes it:

In plain English: every Lightning payment you send costs a flat 0.4%. Receiving is free as long as you already have room. When you don't, Phoenix has to do an on-chain transaction to make room, and you eat the mining fee plus a 1% liquidity fee. Your very first receive also pays a one-time 1,000-sat channel-creation fee.

This is why the top complaint on ACINQ's GitHub is some version of "why am I paying a network fee to receive a Lightning payment?" Because self-custodial Lightning sometimes has to touch the chain, and when the mempool is hot that touch is not cheap. In practice the first funding wants something like 10,000 sats to be worth doing, which hurts on small balances. Phoenix at least refuses to ambush you: it rejects an incoming payment whose fee would top 5,000 sats or half the amount, a cap you can raise, and it lets tiny payments pile up as fee credit instead of forcing a chain operation for every few sats.

So is it a rip-off? No. Is it free? Also no.

Nothing about self-custody is free. You either run your own node and pay in time, hardware, and complexity, or you let ACINQ handle the channels and pay 0.4%. Pick your poison. If you move real amounts and never want to think about liquidity, the fee is a bargain for the headache it removes. If you fire off dust payments all day, the send fee and that first channel will grind on you, and a custodian or your own node is cheaper. Most of the "fee outrage" is really that Phoenix shows you a cost that custodial apps bury in the spread and hope you never notice.

Living with Phoenix: setup and daily use

Setup is where Phoenix earns its reputation. Ben Perrin walked the whole thing end to end in his 2025 BTC Sessions tutorial, and the path is short: write down your 12 words, fund the wallet, let it handle the channel. From there it's one balance you send from and receive to, with the on-chain and Lightning sides stitched together for you.

The daily flow is the payoff. Receiving shows an invoice or a reusable address. Sending scans a QR and shows the fee before it goes. When you want to move savings to cold storage, Phoenix splices funds straight out to an on-chain address, so there's no clumsy withdrawal dance. It shows the fee before each on-chain move and lets you set the feerate, which matters when blocks are full.

The rough edges live at the boundaries. The first surprise on-chain fee on a receive throws new users. And the moment a channel force-closes, the friendly abstraction evaporates and you are suddenly expected to know what a seed sweep is. Phoenix is easy in the normal case and demanding in the rare one, which is the reverse of how it first feels.

The features that matter

ACINQ ships new Lightning standards faster than almost any consumer wallet, because ACINQ helps write them. The parts worth caring about:

• BOLT12 offers. Reusable, non-expiring payment requests that work like a permanent Lightning address. Make one, reuse it forever.
• Splicing. The core trick: one channel resized in place instead of a churn of opens and closes. It's what keeps your balance unified and your long-run fees down.
• Swaproot. Taproot-based swaps that, per ACINQ, make on-chain deposits and withdrawals look like ordinary transactions, cheaper and more private than the old format.
• Taproot channels. Added in late 2025, roughly 15% cheaper per ACINQ and indistinguishable from a normal transaction on-chain.
• Multiple wallets and split PINs. Run several wallets in one app, each with its own Lock PIN and a separate Spending PIN, so someone can receive without being able to spend or see your seed.
• Your own Electrum server, Tor, and a fiat converter round out the control and usability options.

There's also `phoenixd`, a headless server version of the same engine with an HTTP API, built for merchants and developers who want to accept Lightning programmatically without babysitting channels. It is a genuine edge over wallets that stop at the phone.

What to know before you rely on it

This is the part that matters most: what you need to know before you trust it with real money.

• Receiving can cost money, unpredictably. Covered above, still the number-one surprise. Fund in larger, less frequent chunks and budget for it.
• Tor isn't built in anymore. ACINQ ripped out the embedded Tor library in early 2025 for reliability, so routing over Tor now needs a separate app like Orbot, which fights with any other VPN you run. Privacy-minded users are not happy about it, and the friction is real.
• Back up your payment history separately. Restoring from your seed brings back your funds, not your transaction history, unless you exported it (automatic via iCloud on iOS, manual on Android). If your history matters, export the payments database before you need it.
• The builds aren't reproducible yet. Phoenix is open source under Apache 2.0 and its releases are signed, but the build is not yet bit-for-bit reproducible, so nobody outside ACINQ can currently prove the app you downloaded matches the public source. ACINQ admits it and says it's in progress. A known gap, not a security hole, and part of why it's not on F-Droid.
• The "my balance went to zero" panic. Search the forums and you'll find people screaming that a force-closed channel showed a zero balance with no warning. Get the facts straight: your funds are not gone. A force-close settles your money on-chain under your own seed, recoverable after roughly a five-day delay, sometimes through Electrum. It is a genuine UX failure and a real reason to actually understand your seed, but there is no pattern of Phoenix losing user funds, and no Phoenix-specific hack on record.
• You are the backstop. Everything ultimately settles to your seed, so lose the seed and you lose the money, and clawing back a force-closed channel means you have to act. Phoenix is "easy" right up until the day it asks you to be your own bank.

Phoenix vs the alternatives

Phoenix owns one specific seat: the easiest real Lightning self-custody. The others win on different ground.

The honest calls. Zeus is where you graduate when you outgrow Phoenix: point it at your own node and your routing fees fall toward zero, in exchange for real complexity. Wallet of Satoshi is simpler and cheaper for tiny amounts, and it holds your keys, which defeats the entire purpose of holding Bitcoin. Breez's newer wallet dodges the channel minimum by running on the Liquid sidechain, good for small first balances, though you swap trust in ACINQ's node for trust in the Liquid Federation. Aqua is the pick if you want stablecoins next to your Bitcoin, which Phoenix deliberately refuses to touch. Bitcoin-only is a feature, not a limitation. And Muun routes Lightning through submarine swaps rather than a native channel, so its effective fees run higher, which is exactly what sends people to Phoenix.

What you might be looking for instead

• Stacking for the long haul? Lightning is for spending. Savings you're holding for years belong in cold storage, not a hot wallet on your phone.
• Running your own node? Zeus connects to your hardware and drops your fees toward nothing.
• Just buying Bitcoin? A wallet is not an exchange. Strike and friends handle recurring buys, then you sweep the coins to a wallet like Phoenix.
• Want the full menu? Every option lives in the mobile wallets directory.

Who Phoenix is for

What to do this week

If you're new to self-custody: download Phoenix, write your 12 words on paper instead of a screenshot, and fund it with an amount you'd happily carry as walking-around money. Treat the first on-chain fee as the cost of admission.

If you're leaving a custodian: set up Phoenix first, then drain the custodial app to it in one or two larger transfers instead of a dozen small ones, to keep channel and mining costs down. Get your coins off someone else's books.

If you're a US user who bailed in 2024: come home. Reinstall from the App Store or Google Play and restore your seed or start fresh.

If you run your own node: Phoenix is a convenient pocket wallet, not your endgame. Keep Zeus pointed at your node for the cheap, private routing you actually want.

And whatever you do, back up your seed properly today, before a force-close ever forces the lesson. The wallet is easy. Being your own bank is the part that takes work, and it's the part worth doing.

Frequently asked questions