What does it cost to send AQA? Who gets that money? And how does AntiQua stop spam without making everyone pay for it? 
This post explains the AntiQua fee system exactly as it works in the node software today โ no marketing numbers, just the rules the network enforces. Every number below was re-checked against the NodeCore source in September 2026.
TL;DR
A tiny base fee + your tip. Every transaction pays a small, fixed minimum of 100 satoshi per KiB โ a typical post-quantum payment about 0.00001 AQA. On top, you choose a tip: a higher tip gets you confirmed faster.
Every fee goes to the miner. Nothing is burned, nothing goes to a hidden account.
More inputs cost more โ because every input carries a large post-quantum signature.
Spam is stopped with work, not money: from the 6th pending transaction, your wallet has to solve a tiny proof-of-work.
Everything is exact: all amounts are whole satoshi โ no rounding, no floating point.
What a fee actually is
On AntiQua, the fee isn't a separate field you can fake โ it's simply what's left over:
fee = sum of inputs โ sum of outputs
Everything you don't send to a recipient or back to yourself as change is the fee, calculated to the exact satoshi (1 AQA = 100,000,000 satoshi). The node checks that this amount covers everything the transaction requires โ otherwise it's rejected.
The fee is also part of the transaction hash: if you change the fee, you get a new transaction ID. Nobody can alter it on the way.
Base fee + your tip
Your fee has two parts:
The base fee โ a fixed minimum set by consensus: 100 satoshi (0.000001 AQA) per started KiB of transaction size. "Started" is literal: 9,194 bytes round up to 9 KiB. A typical post-quantum payment is around 9โ10 KiB, so the base fee is roughly 900โ1,000 satoshi โ about 0.00001 AQA. It applies to every transaction from block 1 on.
The tip โ whatever you add on top for the miner:
Low tip โ still valid (even a tip of 0 is accepted, as long as the base fee is covered; the dust limit and spam protection still apply), but you may wait longer when blocks are busy
Higher tip โ miners pick your transaction first
Fixed, not adjustable: the base fee depends only on the block height โ no node, miner or admin can raise it on the fly or tune it to its own mempool. The consensus code allows a ceiling of 1,000ร the minimum, but the active value is the minimum, and nothing in the node moves it. Changing it would require a network upgrade, announced in the Core Changelog first.
Real example from our testnet milestone: sending 10 AQA with a 0.01 AQA tip โ the miner collected the fees on top of the block reward, the receiver got exactly 10 AQA. (See the milestone post) That payment was made before the base fee went live โ today the wallet would add about 0.00001 AQA on top.
What your wallet suggests
You don't calculate any of this by hand. The wallet adds the base fee automatically (it comes out of your change) and offers three lanes:
| Lane |
What it does |
Economy |
No extra tip at all โ just the base fee. Fine when you're not in a hurry. |
Recommended |
The median tip per KiB that the network is currently paying, never less than the base rate. |
Next block |
The 90th percentile of current tips, at least one step above Recommended. |
When the wallet has seen no recent traffic โ an empty or fresh mempool โ it falls back to the base rate for Recommended and twice that for Next block.
These lanes are wallet policy, not consensus. Another wallet may suggest different numbers; the network only enforces the base fee.
What happened to the relay fee?
Older wallet builds showed a third field called relay fee, and our earlier testnet posts mention it. It no longer has any role. The node requires zero for relay: the amount a transaction must cover is base fee + tip (+ contract deployment fee), nothing else.
The name always caused the same misunderstanding, so to be clear: a relay fee never went to relay nodes. Relay nodes are paid only through the Relay Reward Lottery โ 1.5 AQA per block from its own 4 million AQA budget, never out of your transaction.
In practice: if your wallet still shows the field, leave it at 0 and use the tip. The tip is strictly better anyway: both raise your fee rate equally, but only the tip counts as the tie-breaker when two transactions have the same score.
Footnote: once the lottery budget is used up (in roughly 50 years), 5 % of all fees in a block go to the lottery winner. That's a share of the block's total fees โ not "the relay fee goes to relays".
Why more inputs cost more
Every coin you spend (an input) has to be signed โ and post-quantum signatures are big, several kilobytes each. A transaction that gathers many small coins is heavy for every node on the network. So the required fee grows with the number of inputs:
| Inputs |
Fee factor |
| 1 |
1.00ร |
| 2 |
1.05ร |
| 3 |
1.30ร |
| 4 |
1.60ร |
| 5 or more |
2.00ร |
The factor applies to base fee and tip together, and it also lowers a transaction's priority in the mempool โ it sits in the divisor of your fee rate, so a heavy transaction ranks below a light one paying the same absolute amount.
Practical tip: if your wallet has lots of tiny coins, consolidate them once in a quiet moment. At five inputs you are paying double the floor on every send.
Change that disappears into the tip
Sometimes a payment leaves a few hundred satoshi over. Your wallet will not hand that back as change, and that is deliberate:
- Leftover below 1,000 satoshi always becomes extra tip. An output that small would be invalid anyway โ see the dust limit.
- Leftover up to 50,000 satoshi may optionally be folded into the tip instead of creating a new coin, to keep your wallet from filling up with crumbs that cost more to spend later than they are worth.
- Anything larger comes back to you as a normal change output.
So if the fee on your receipt looks slightly higher than the lane you picked: that is your own leftover change, not a surcharge.
Who gets confirmed first
Miners sort waiting transactions by a simple score:
- Fee per KiB โ how much you pay relative to the size of your transaction
- Divided by the input factor โ heavy transactions rank lower
- Contract deployments get a small boost (see below)
On a tie: the higher tip wins, then the older transaction, then the smaller hash. Completely deterministic โ no favourites.
The mempool holds up to 50,000 transactions or 500 MiB, whichever comes first.
What if I pay too little?
Then the transaction never enters the mempool, and a transaction that is not in the mempool is never passed on to anyone. You get an error straight away โ FEE_TOO_LOW if the amount misses the policy floor, FEE_INSUFFICIENT if it covers the floor but not everything the transaction owes.
There is no second, lower "minimum relay fee" hiding behind this one. What the floor says is what the network takes.
And no, a light node isn't a way around it. A transaction gossiped in from a light client runs through exactly the same check as any other โ the light path only serialises the fee fields, it never applies a floor of its own.
Where the money goes
Every fee in a block goes to the miner who found it:
Miner reward = block reward (50 AQA, halving) + all fees in the block
Nothing is burned. Fees don't disappear, they pay the people securing the network.
Relay nodes are paid separately โ through the Relay Reward Lottery, 1.5 AQA per block from its own 4 million budget, not by taking a cut of your fee. (Details in the tokenomics post)
In the far future โ once the lottery budget is used up after roughly 50 years โ 5 % of the block fees go to the relay winner, the rest stays with the miner.
Spam protection without making everyone pay
Making fees high enough to stop spam would make the network expensive for everyone. AntiQua keeps the base fee tiny and adds a separate limit instead:
1. Micro proof-of-work 
- Your first 5 pending transactions are completely normal.
- From the 6th pending transaction (from the same sender), each one must carry a small proof-of-work โ its hash needs 14 leading zero bits. For a normal user that's a blink; for someone trying to flood the network with thousands of transactions, it adds up fast.
2. A hard ceiling per sender 
- The same owner can have at most 25 unconfirmed transactions waiting at once. Number 26 is refused with
QUOTA_EXCEEDED until some of them confirm.
- That is the real admission limit, and it's the same number in the mempool and in the node's submit check. If you have read about 100, 200 or 1,000 anywhere: those numbers exist in older material but no code path enforces them.
Both are node policy, not consensus: they govern what a node accepts into its own mempool and passes on.
Plus the dust limit: outputs smaller than 0.00001 AQA are invalid, so nobody can clog the chain with worthless crumbs. How the dust limit works
Smart contract deployment fees
Deploying a smart contract adds a small, fixed deployment fee on top โ depending on the template:
| Template |
Base deployment fee |
| Escrow |
0.0001 AQA |
| Time-Lock |
0.00015 AQA |
| Token ยท NFT ยท Vesting |
0.0002 AQA |
| Voting |
0.00025 AQA |
| Recurring payment |
0.0003 AQA |
The deployment fee can rise up to 5ร when the chain is busy โ calculated only from on-chain data (block timing and number of contracts), so every node arrives at exactly the same number. Deployments also get a 15โ25 % priority boost in the mempool, so contracts don't get stuck. Calling a contract later costs a normal transaction fee.
The rules in one place
- Fee = inputs โ outputs, exact to the satoshi, part of the transaction hash
- Base fee 100 sat per KiB (consensus, from block 1) + your tip. There is no relay fee any more โ the node requires zero for relay
- More inputs โ higher fee factor and lower priority; the factor hits base fee and tip together
- Leftover change below 1,000 satoshi always becomes tip
- 100 % of fees to the miner, nothing burned โ relays are paid by the lottery
- Micro proof-of-work from the 6th pending transaction; at most 25 unconfirmed per sender
- Contract deployments: small fixed fee by template, up to 5ร under load
Questions?
Wondering what fee to set, or how this compares to Bitcoin or Ethereum? Ask below โ happy to explain. 
โ Blythex