Pushpendra Singh TxFlow review: seven months on a brand-new perp DEX
A perpetuals DEX on its own L1, live since March 2026. Non-custodial, gated, and copying Hyperliquid's fee card exactly. What it has not proved yet.
I have had an account on TxFlow for a while now, traded on it, and withdrawn from it. This is what I actually think, including the parts that will cost me referral income.
What it is
TxFlow is an exchange that runs on its own Layer 1 chain. It launched on 28 March 2026, so in November 2026 it is roughly seven months old. The founders are listed as Matt Hu, Harrison Rea and Jack Wang, and the project describes itself as community-owned rather than company-controlled.
The product is perpetual futures first: around 150 perpetual pairs against roughly 20 spot pairs. Collateral is USDC. It is non-custodial, so there is no exchange balance, and no KYC, though there is an access gate I will come back to.
The one design decision that matters
Most DEXs you have used are automated market makers. You trade against a pool and the price slides along a curve.
TxFlow runs a central limit order book, fully on-chain. Same structure as Binance or Bybit: real limit orders, real market orders, a visible book with depth on both sides. For perpetuals this is not a preference, it is close to a requirement. An AMM cannot give you a tight fill on a leveraged position, which is why on-chain perps took so long to become usable.
If you have avoided decentralised exchanges because they felt nothing like the trading screens you know, this will feel familiar in a way those did not. That is the genuine achievement here, and I do not want the criticism below to drown it.
Fees, and the claim you should stop repeating
Base tier is 0.015% maker and 0.045% taker on perpetuals, 0.040% and 0.070% on spot. Tiers run VIP 0 to VIP 6 on 14-day volume, with maker fees reaching 0% above $500 million and top-tier taker at 0.024%. Deposits are free. Withdrawals cost 0.1 to 0.5 USDC by network.
Those are good rates, genuinely below most centralised venues.
They are also identical to Hyperliquid’s, digit for digit. TxFlow has adopted that fee card wholesale. So “lower fees than Hyperliquid” appears on a great many TxFlow pages and is simply untrue, and Hyperliquid can be the cheaper venue because staking HYPE cuts its fees a further 5% to 40% with no TxFlow equivalent. Full side-by-side in TxFlow vs Hyperliquid.
What I like
Tokenised stocks and commodities on the same collateral. US equity, metals and oil futures beside crypto perps, from one USDC balance, no brokerage account, no separate onboarding. That is a real differentiator rather than a repackaged one. Whether demand for it lasts is open, but nothing else on my partner list does it.
Non-custodial, and it means something. After the year Indian users have had with exchange incidents, including CoinDCX confirming a $16 million loss in October 2026 after first denying it, an architecture where the venue never holds your coins is not a talking point. It removes an entire category of risk.
The fill quality on major pairs is better than I expected for something this young. On BTC and ETH perps at ordinary retail size I have not had a fill that made me wince.
What I do not like
Seven months is a sample, not a track record. It has not traded through a violent move yet. Every venue looks solid until the day it does not, and the information you most want about an exchange is its behaviour on the worst day of the year. That data does not exist here.
Audit coverage is the bridge only. OpenZeppelin audited the bridge contract. Auditing the bridge first is the right instinct, because bridges are where the largest losses in this industry have happened. But it is one contract, and “audited” in marketing rarely distinguishes between some and all.
The 250,000 transactions per second figure is self-published. It appears in nearly every article about TxFlow and is always sourced to TxFlow. It may be accurate. It is not evidence, and the uniformity with which it is repeated tells you how much independent checking has gone into the coverage.
Liquidity is thinner than the majors, and on a perp venue that is the weakness you cannot argue around. A 0.045% taker fee on a $10,000 position is $4.50. Filling half a percent worse than the screen costs $50. Slippage dwarfs fees, it scales with your size, and it never appears in a fee comparison.
The access gate flatters the numbers. TxFlow needs an access code from its Discord. Gating filters for intent and makes early metrics look tidier per user than open access would. You find out what the real demand was on the day the gate comes off. Practical guide to getting through it in how to get a TxFlow access code.
The warning I would put above all of this
TxFlow’s own documentation advises against account-abstraction wallets (smart contract wallets rather than standard key pairs) and states the worst-case outcome is funds stuck in the contract.
I read a dozen TxFlow reviews while researching this. One mentioned it.
If your wallet recovers from a 12 or 24 word seed phrase, you are fine. If you sign in with email or a social account, or recovery runs through guardians, check before depositing. A hardware wallet is in the safe category here, not the risky one. Ninety-second check in the wallet that gets your funds stuck.
On the airdrop
Most people asking me about TxFlow are asking whether farming it now pays later.
There is no confirmed airdrop. A points system and a fee rebate, and nothing promised. Several review sites imply a retroactive distribution, and that implication does a lot of work in their referral funnels.
Farming an unconfirmed airdrop means paying certain fees today, on volume you would not otherwise trade, for a distribution nobody has promised at a size nobody has stated on a date nobody has set. Sometimes it pays spectacularly. The failures are invisible because nobody screenshots the protocol they farmed that never distributed.
Who it is for
Use it if you want tokenised equity or commodity exposure alongside crypto perps, you already understand leverage, you trade at a size where thin books do not hurt you, and you can treat the position as speculative.
Do not use it if you are new to perpetuals, you need depth for large orders, you hold a smart contract wallet, or the airdrop is your only reason.
How much? Less than feels reasonable. Non-custodial removes the risk of the venue losing your balance. It does not remove smart contract risk, bridge risk or chain risk. It moves the failure mode, it does not delete it.
My own position is small enough that if the chain stopped tomorrow it would annoy me rather than hurt me. That is not a prediction about TxFlow, it is the only honest answer to a question nobody can answer at seven months, and I wrote about why I stopped making the other kind of call in why I stopped giving price predictions.
Longer assessment of the hype question, with the four numbers that will settle it, in will TxFlow be the next hype.
txflowdexperpetualsreviewno-kyc
Common questions
What is TxFlow?
A perpetuals and spot exchange running on its own Layer 1 chain, with a fully on-chain central limit order book. It launched on 28 March 2026, is non-custodial, requires no KYC, and uses USDC as collateral. It also offers tokenised US stock, metals and oil futures alongside crypto perpetuals.
Is TxFlow legit?
It is a real, functioning exchange rather than a scam, and it is non-custodial so it never holds your balance. But it is only seven months old, only its bridge contract has a public audit, and its headline performance figures are self-published. Legitimate and proven are different things, and TxFlow is the first but not yet the second.
What are TxFlow's fees?
Base tier is 0.015% maker and 0.045% taker on perpetuals, and 0.040% and 0.070% on spot. Volume tiers run from VIP 0 to VIP 6, where maker fees reach zero above $500 million of 14-day volume. These rates are identical to Hyperliquid's.
Who should not use TxFlow?
Anyone new to perpetuals, anyone who needs deep liquidity for large orders, anyone using a smart contract wallet, and anyone whose only reason for being there is an airdrop that has never been confirmed.