This review contains affiliate links. If you buy via these links, I may earn a commission at no extra cost to you. It doesn’t change what I say about the product.
TL;DR
Phoenix Wallet is the Bitcoin-only Lightning mobile standard. It is self-custodial — you hold a 12 or 24-word seed, your keys are on your device — while auto-managing Lightning channels so you never have to think about inbound liquidity or channel rebalancing. ACINQ, the Paris-based company behind Phoenix, is one of the most technically credible teams building on Lightning, and that expertise shows in every detail of the product. If you want self-custodial Lightning payments on your phone without running a full node, Phoenix is the answer.
What Phoenix is
ACINQ is not primarily a wallet company. It is a Lightning infrastructure company that also ships a consumer wallet. The company has built Eclair — one of the three dominant Lightning Network node implementations (alongside Core Lightning and LND) — and maintains phoenixd, the server-side counterpart to the mobile Phoenix wallet. The company has been actively contributing to the Lightning protocol specification since 2016.
This background matters because Phoenix is not a startup’s first Lightning experiment. It is a consumer product built by the team that writes one of the protocol implementations, contributes to the specification, and runs significant routing node infrastructure on mainnet. When Phoenix adds BOLT 12 support or implements channel splicing, ACINQ is often the team that helped write those specs.
Phoenix is available on Android (via Google Play and direct APK) and iOS (via the App Store). The source code is open under the Apache 2.0 license on GitHub at github.com/ACINQ/phoenix. There is no desktop application, no web app, and no browser extension — Phoenix is a mobile-only product, deliberately so.
Setup experience
Opening Phoenix for the first time is almost refreshingly simple for a self-custodial Bitcoin wallet. There is no account creation, no email address, no KYC verification of any kind. You are presented with a choice: create a new wallet or restore an existing one.
Choosing “create a new wallet” generates a 12-word seed phrase (Phoenix defaults to 12 words for the standard seed, with an option to use a BIP-39 passphrase for an additional factor). You are prompted to write it down and confirm several words. The seed is generated on-device. ACINQ never sees it. This is the master key to your funds — lose it without a backup and the funds are gone.
The first time you receive Bitcoin on Lightning, Phoenix needs to open a channel to the ACINQ routing node. This is where new users encounter the first friction point: the channel-opening fee.
When you receive your first Lightning payment, Phoenix automatically opens a channel on your behalf. This involves an on-chain transaction, which costs mining fees. Phoenix charges a service fee for this operation — as of early 2026, the fee structure is a combination of a mining fee pass-through plus a small fixed ACINQ service fee. For a first payment of, say, 10,000 sats, a channel-opening fee of 3,000 sats is a significant percentage overhead. For a first payment of 100,000 sats, the same fee becomes negligible.
This fee model confuses many new users who expect to receive exactly what was sent. It is worth understanding that the fee is not a scam or an error — it is the cost of having Phoenix open and manage a Lightning channel automatically on your behalf, paid to the miners and to ACINQ for the service. Once the channel is open, subsequent payments have no channel-opening fee until the channel needs to be resized.
Phoenix handles this transparently: before any fee-incurring operation, it shows you the exact fee and asks for confirmation. You are never surprised. But you do need to understand what you are confirming.
Day-to-day UX
Once you have a funded channel, Phoenix’s day-to-day experience is among the smoothest in the Lightning wallet space.
Sending: You paste a Lightning invoice (BOLT 11) or scan a QR code, confirm the amount and fee, and the payment goes through. For most payments on well-connected nodes, Phoenix routes successfully on the first attempt. The routing is handled by ACINQ’s infrastructure, which runs a significant part of the Lightning routing network. Failure rates are low.
Receiving: Share your Lightning address (in the format name@domain if you have a Lightning address configured, or via a generated invoice) and payments arrive. If the incoming payment is larger than your channel can accept, Phoenix automatically manages a splice-in to resize the channel — with your confirmation and a fee.
BOLT 12 offers: Phoenix supports BOLT 12, the newer Lightning payment format that allows static, reusable “offers” (analogous to a Lightning address, but protocol-native). A BOLT 12 offer is a string that someone can send to you repeatedly without generating a new invoice each time. This is particularly useful for receiving recurring payments, tips, or donations. BOLT 12 also supports improved privacy through blinded paths — the payer cannot trace the exact routing path to your node.
Swap-in and swap-out: Phoenix allows you to swap on-chain Bitcoin into Lightning (adding to your channel balance) and swap Lightning balance out to an on-chain address. These operations involve fees — a combination of on-chain mining fees and a small ACINQ service fee — but they make Phoenix unusually flexible as a single wallet that handles both layers.
Auto-liquidity: When you need more inbound liquidity than your current channel provides, Phoenix can splice in additional capacity automatically. This is one of Phoenix’s most significant UX innovations: the channel is never “full” in the sense of blocking payments. It grows to accommodate what you need, with fees you confirm.
Self-custody model
The defining feature of Phoenix, compared to custodial Lightning wallets, is that you hold your own seed.
Phoenix implements a non-custodial channel model using a construct ACINQ calls “self-custody Lightning channels.” In technical terms, you and ACINQ share a 2-of-2 multisig channel. Your keys sign outgoing transactions. ACINQ cannot move your funds unilaterally because their signature alone is not sufficient.
The seed phrase — 12 words by default, 24 words if you choose the extended option — encodes your private keys in standard BIP-39 format. Phoenix also supports bLIP-39, a proposed extension that encodes additional channel state in the seed phrase derivation, enabling more complete channel recovery from seed alone. This means that if you lose your device, you can reinstall Phoenix on a new phone, restore your seed, and recover both your on-chain funds and your Lightning channel state.
The recovery from seed is not instantaneous — ACINQ needs to cooperate in the recovery process to close channels and return funds — but the keys are yours and the funds cannot be taken without your cooperation. This is meaningfully different from custodial wallets where the provider can freeze, block, or confiscate funds at will.
You can export your seed from Phoenix at any time in Settings. You can verify the seed. You can import it into compatible Lightning wallets. The transparency here is genuine: Phoenix does not lock you to ACINQ’s infrastructure.
That said, the channel management is not entirely trustless. Phoenix’s auto-channel UX means ACINQ’s node is the channel peer for all Phoenix channels. This is a routing trust dependency: Phoenix routes through ACINQ’s infrastructure. A competent attacker who could intercept or analyse ACINQ’s Lightning node traffic could infer payment patterns, though not steal funds. For users with privacy-critical use cases, this routing centralization is a consideration.
Trust model: Phoenix vs full-node-plus-Zeus
The right comparison is not Phoenix versus custodial Lightning — Phoenix clearly wins that comparison. The more interesting comparison is Phoenix versus running your own Lightning node (such as Umbrel or Start9 running LND) paired with Zeus Wallet, which connects to your own node.
Phoenix strengths: No server to manage. No uptime responsibility. Works from a fresh phone in minutes. Automatic channel management. Lower technical bar. The UX is genuinely excellent.
Phoenix trade-offs: Your channel peer is always ACINQ’s node, not a node you control. You cannot see or control the raw channel state. If ACINQ’s routing infrastructure has issues, your payments may fail even if your keys and funds are secure. You are trusting ACINQ to cooperate in channel recovery.
Full-node-plus-Zeus strengths: You control the routing peer (your own node). You can inspect the complete channel database. No dependency on any third party’s infrastructure for routing. For the most privacy-sensitive users, the routing path from your own node reveals less about the payer than routing from an ACINQ-operated Phoenix channel.
Full-node-plus-Zeus trade-offs: Managing a Lightning node requires maintaining an always-on server, monitoring channel health, managing inbound liquidity yourself, and handling channel failures manually. This is meaningful operational work. For most people who aren’t specifically running this for a business or as a Bitcoin hobby, it’s overkill.
Phoenix sits in a practical sweet spot: materially less trusted than a self-hosted node, but massively more practical than one for everyday Lightning use. For the person who wants Lightning self-custody without the node management burden, Phoenix is the correct tool.
Phoenix vs alternatives
Muun: Muun historically occupied a similar niche to Phoenix — self-custodial Lightning on mobile. Muun has since shifted its model in ways that make it less comparable as a pure Lightning wallet. Muun also uses a submarine-swap approach (every Lightning payment is ultimately an on-chain swap) rather than native channels, which results in different fee structures and confirmation times. Phoenix’s native channel approach is generally faster and cheaper for established users.
Wallet of Satoshi: Custodial. Not comparable for self-custody users. WoS is convenient but Wallet of Satoshi holds your keys. The company has demonstrated its willingness to cut off users in certain jurisdictions. Phoenix users have sovereign custody; WoS users do not.
Zeus: Zeus is a power-user wallet designed to connect to your own Lightning node (LND, Core Lightning, Eclair) or to use a built-in node for advanced users. Zeus is more capable than Phoenix in the sense of exposing more control and supporting more node operations, but it requires significantly more technical knowledge to operate correctly. If you want to run your own node, Zeus is the front-end for it. If you don’t want to run a node, Phoenix is the product.
Strike, Cash App Lightning: These services offer Lightning payments but are custodial and KYC-required. Not comparable.
Who it’s for — and who it’s not for
Phoenix is ideal for:
The daily Lightning spender who wants self-custody. If you’re using Lightning to pay for coffee, split bills with friends, send tips or donations, receive payment for freelance work in Bitcoin, or simply maintain a mobile Lightning wallet for everyday spending, Phoenix is the best product in this category. The automatic channel management means you never need to think about channel state, and the BOLT 12 support future-proofs your payment addresses.
People stepping up from custodial Lightning. If you currently use Wallet of Satoshi, Strike, or any other custodial Lightning service and want to move toward self-custody without the complexity of running a node, Phoenix is the natural upgrade path.
Bitcoiners who travel and want a reliable Lightning wallet across jurisdictions. No KYC means no jurisdiction-specific account issues.
Phoenix is not ideal for:
People who need full node sovereignty. If your threat model requires that you cannot depend on any third party’s routing infrastructure — for example, journalists in repressive regimes, or users with significant adversarial pressure — you need a full self-hosted node. Phoenix’s routing trust in ACINQ’s infrastructure is a real limitation here.
Large Lightning balances. Lightning channels are not the right place for savings-tier Bitcoin. Channel management has costs and risks — a remote force-close, a routing node going offline, a protocol-level issue — that are not appropriate for money you cannot afford to lose temporarily. Use hardware cold storage (a Coldcard or Trezor) for savings. Use Phoenix for spending.
Power users who want to inspect and control their own channel state. If you’re the kind of user who opens htlcinterceptors and inspects channel databases, you want Zeus on your own node. Phoenix abstracts all of this away, which is a feature for most users but a limitation for you.
The fee structure in practice
Phoenix’s fee structure deserves its own clear explanation, because it confuses users who expect Lightning to always be zero-fee.
There are three types of fees you will encounter:
Channel-opening fees: Paid once when a channel is opened or spliced. Covers mining fees for the on-chain transaction plus a small ACINQ service component. Unavoidable for the first time you receive Lightning, or if your channel needs more capacity.
Routing fees: Paid to Lightning routing nodes when Phoenix routes a payment. Typically small — a few satoshis per payment for most amounts — and displayed before you confirm. ACINQ takes a small share as the routing infrastructure provider.
Mining fees for on-chain operations: If you do a swap-out (moving Lightning balance to on-chain Bitcoin), you pay the prevailing mempool fee. Phoenix uses fee estimation to keep this reasonable, but during high-fee environments this can be material.
The honest assessment: Phoenix is not the cheapest Lightning wallet for tiny amounts. A 1,000-sat payment to a fresh wallet with no existing channel will pay proportionally high channel-opening fees. Phoenix is economical at scale — regular payments, established channels, avoiding unnecessary swap operations. The fee model rewards users who treat Phoenix as a regular payment wallet, not a one-off-use tool.
Verdict
Phoenix earns its 4.7 rating without qualification. The 0.3 deduction is almost entirely for the channel-opening fee UX — a friction point that is technically necessary but frequently confusing to new users who expect a flat-fee or zero-fee experience.
The case for Phoenix is simple: if you want self-custodial Lightning on your phone, Phoenix is the best implementation of that idea on the market. ACINQ has done the hard work of making Lightning channels manageable for non-technical users while preserving the security property that actually matters — you hold your keys, and those keys are yours. The BOLT 12 support, the splicing mechanism, the bLIP-39 seed recovery, the open-source codebase — all of these represent genuine technical depth that you will not find packaged this cleanly in any other mobile Lightning wallet.
Bitcoin-only. Open source. Self-custodial. No KYC. Leading protocol support. Phoenix ticks every box on the checklist that matters. Install it.
Sources:
- Phoenix official site: phoenix.acinq.co
- ACINQ GitHub (Phoenix source code, Apache 2.0): github.com/ACINQ/phoenix
- Eclair Lightning implementation: github.com/ACINQ/eclair
- BOLT 12 offers specification: github.com/lightning/bolts/pull/798
- bLIP-39 seed backup proposal: github.com/lightning/blips/blob/master/blip-0039.md
- phoenixd (server-side Phoenix): phoenix.acinq.co/server
Related reading
- Is BOLT12 safe to use yet? — Phoenix’s BOLT12 support is one of the strongest in the ecosystem; this Q&A walks through the protocol’s status across other wallets if you need a recipient that doesn’t run Phoenix.
- Lightning address verifier — for verifying any LNURL / Lightning offer / static QR you receive before sending.
- BOLT-12 decoder and BOLT-11 decoder — inspect the underlying invoice payload byte-by-byte.