Paid pool acceleration works, free rebroadcasting usually does nothing, and most people who pay did not need to. In the records we sampled, the median accelerated transaction was already paying 2.2 sat/vB, comfortably enough for the next block. Quotes ran 3 to 17 times the real fee shortfall.
Ask this question anywhere and you get two answers, both useless. Accelerator sites say yes, obviously, here is our button. Reddit says no, it is a scam, just wait. Neither camp shows any numbers.
So we went and got some. We queried the public records of the largest pool backed accelerator, pulled its live pricing for real unconfirmed transactions, recomputed the hashrate its partners actually control from block data, and checked which competing services are even alive. This article is what came back, and it is re-measured every morning.
The honest summary is that the technology is real and the pricing is not. There is a genuine mechanism by which a mining pool can push your transaction into a block, and it does work. There is also a large gap between what that service is worth and what it is sold for, and a much larger population of people buying it who never needed it.
We run a free accelerator too. It rebroadcasts to twenty independent nodes and it tells you when that will not help you, which is most of the time. That is the only reason we are comfortable publishing the numbers below.
What people actually mean by an accelerator
The word covers four completely different things that share a marketing category and nothing else. Getting them straight is most of the work, because arguments about whether accelerators work are usually two people describing different mechanisms.
| Mechanism | What it does | Who can use it | Does it work |
|---|---|---|---|
| Free rebroadcast | Resends your transaction to well connected nodes | Anyone | Only if the network lost it |
| Paid pool acceleration | Pools treat your transaction as higher paying when building blocks | Anyone who pays | Yes, weighted by their hashrate |
| RBF | You re-sign the same payment with a higher fee | Sender only | Yes, and it is nearly free |
| CPFP | A high fee child transaction drags the parent into a block | Sender or receiver | Yes |
For per service quotes on a single live transaction, see paid Bitcoin accelerator reviews.
Only the first two are what people buy. The last two are what most people should be doing instead, and they have their own guides: Replace-By-Fee and Child Pays For Parent.
What a free rebroadcast really does
A free accelerator takes your transaction ID, fetches the raw transaction, and pushes it to a list of nodes. Ours pushes to twenty. Others claim ten or twenty. The list is not the interesting part.
The interesting part is what happens when those nodes already have your transaction, which they almost always do. Bitcoin gossips aggressively. A transaction broadcast from a normal wallet reaches most of the network within seconds. When we relay something that has been sitting unconfirmed for hours, most nodes answer with some version of already have that one.
Sending it again to a node that already holds it changes nothing. It does not raise the fee, it does not move the transaction up any queue, and it does not make a miner more likely to select it. Miners choose transactions by fee rate, and rebroadcasting does not touch the fee rate.
The claim to be suspicious of
If a site tells you rebroadcasting will speed up a transaction the whole network already holds, that claim is false. It is the most common piece of misinformation in this niche, and it is what makes people dismiss the entire category as a scam.
Why the free tier still exists
Because it is cheap to run and it converts. A visitor arrives worried, presses a button, watches a list of green ticks appear, and feels helped. Nothing happened on the network, but the visitor is now on a page with ads on it, or looking at a paid upgrade.
That is the honest explanation for why nearly every free accelerator shows a wall of identical success badges and no detail. Detail would reveal that most of those nodes already had your transaction.
The one case where rebroadcasting is the cure
There is a real scenario, and it is why we still run the free service. If the network genuinely does not have your transaction, pushing it out is exactly right. That happens when a broadcast failed silently, or when the transaction was dropped after the fourteen day expiry. The guide on vanished transactions covers both cases and how to tell which you have.
Find out which case you are in
Paste your transaction ID and we tell you whether the network has it, what it is paying, and whether rebroadcasting will do anything for you. Free, no signup.
How paid pool acceleration works
This one is real, and the mechanism explains both why it works and why developers dislike it.
When a mining pool builds a candidate block it picks transactions from its own mempool, sorted by fee rate, until the block is full. An accelerator has an agreement with a set of pools. You pay the accelerator, the accelerator tells those pools to treat your transaction as though it paid a much higher fee, and the pools slot it in accordingly.
Nothing about your transaction changes. The fee written into it stays the same, the transaction ID stays the same, and every other node still sees the original low paying transaction. The extra money moves off chain, from you to the accelerator to the pool. This is why the practice is called an out of band fee.
Why it is probabilistic, not guaranteed
The pools you paid only control part of the network's hashrate. They can only include your transaction in blocks they themselves find. If the next block is found by a pool outside the arrangement, your transaction is not in it, and you wait for the next one.
So the real promise is not next block. It is included in the next block found by one of these pools, which is a probability rather than a schedule.
We checked the hashrate claim
The largest accelerator says its partner pools control over 80 percent of hashrate. That claim gets repeated by resellers without verification, so we tested it. We took the list of pool identifiers the service returns, matched them against the pools that actually found blocks over the past month, and added up the blocks.
The claim holds up. Foundry USA, AntPool, F2Pool, SpiderPool and MARA between them account for most of it. Notably ViaBTC, which runs its own competing accelerator, is not in the list, and neither is OCEAN.
What the public acceleration records show
The largest accelerator publishes its own history through an open endpoint, and almost nobody reads it. We do, every morning.
| Metric | Value | What it tells you |
|---|---|---|
| Accelerations requested | 30,381 | Total since April 2023 |
| Accelerations completed | 29,775 | Completion rate 98 percent |
| Total paid to miners | 2.909 BTC | About $226,046 over three years |
| Recent month | 876 | About 29 per day now |
Two things jump out. The whole market is small: fewer than thirty accelerations a day, worldwide, through the dominant provider. And it has shrunk by roughly three quarters since late 2024, which lines up exactly with fees collapsing and staying collapsed.
The completion rate looks fantastic until you think about what it measures. It counts transactions that confirmed after being accelerated. In a market where almost everything confirms quickly anyway, that number would be high whether the service did anything or not.
The detail that changes the picture
We sample the individual acceleration records daily. For each one we can see the fee rate the transaction was already paying before anyone paid anything.
What the sampled accelerations were actually paying
- Median fee rate before acceleration: 2.2 sat/vB
- Median extra paid to miners: 2,000 satoshis
- Median time from paying to confirmation: 18.3 minutes, roughly two blocks
Read the middle line again. The median customer paid, then waited about two blocks. Not next block.
And the first line is the uncomfortable one. A transaction paying above the current clearing rate was already going to be mined. Somebody paid to accelerate a transaction that needed no help at all. We have seen people accelerate transactions paying over 11 sat/vB, several times the going rate, in a market where nothing anywhere pays more than about 6.
What it costs against what it is worth
The pricing endpoint is public too, so you can ask what a specific transaction would cost to accelerate and compare that against the actual gap between what it pays and what the next block needs.
| Transaction | Actual fee shortfall | Cheapest option offered | Markup | Most expensive option |
|---|---|---|---|---|
| 2,660 vB, 19 ancestors | 1,729 sats, about $1.34 | 10,000 sats, about $7.76 | 5.8x | 30,000 sats (17.4x) |
| 3,162 vB, 25 ancestors | 5,672 sats, about $4.40 | 17,500 sats, about $13.58 | 3.1x | 60,000 sats (10.6x) |
The shortfall column is the honest number. It is what the transaction is missing to reach the next block, and it is what you would pay a miner if you simply raised the fee yourself.
To be fair to the provider, some of that gap is a real business. They maintain pool relationships, run infrastructure, take payment risk, and are entitled to a margin. The question is whether the margin is 3x or 17x, and whether the buyer has any way to see which they are getting. Currently they do not, because the shortfall is never shown next to the price.
Why so many paid accelerations were unnecessary
This comes down to a change in the fee market that most people have not registered, and that most accelerator marketing still ignores.
The mempool holds 85,351 transactions and looks 44 blocks deep. That sounds like severe congestion, and screenshots of it get used to sell acceleration. But only 2.12 percent of that backlog pays 1 sat/vB or more, and 69 percent of it sits between 0.1 and 0.2 sat/vB.
So the picture an accelerator shows you is technically accurate and completely misleading. We take it apart properly in why the mempool is full but fees are 1 sat/vB.
Why Bitcoin developers dislike the practice
This rarely appears on accelerator sites, and it is the strongest argument against routine use.
Bitcoin Optech, the technical newsletter most protocol developers read, describes out of band fees as a censorship resistance problem. The reasoning is simple. Miners with more hashrate find blocks more consistently, so a user who wants speed will pay the biggest miners, because paying a small pool buys almost nothing. If paying miners directly becomes normal, fee revenue concentrates in the largest pools, and the number of parties you would need to compromise in order to censor a transaction gets smaller.
There is a second, subtler cost. Fees paid inside a transaction are visible to everyone and inform fee estimation for the whole network. Fees paid off chain are invisible. The more acceleration happens, the less anyone can tell what block space actually costs.
When paying is genuinely justified
There is a real case, narrower than the marketing suggests.
You cannot raise the fee yourself
This is the main one, and it almost always means an exchange withdrawal. The exchange signed the transaction, so you cannot replace it. The exchange guide covers this in full.
The receiver cannot help either
A CPFP from the receiving side solves many of these situations for a few hundred satoshis. If the recipient is a service that will not do that, acceleration is back on the table.
The delay is genuinely expensive
If a payment window closes or a counterparty walks, spending a few dollars to raise the odds on the next block is a rational trade. Most delays are not like this. Most delays are somebody refreshing an explorer at two in the morning.
The test we would apply
Pay only when all three are true. You cannot raise the fee yourself, nobody on the other side can either, and waiting another few hours would cost you more than the quote. If any one is false, keep your money.
How to check before you pay anyone
- Confirm it is really unconfirmed. A surprising share of panics are about transactions that already confirmed.
- Find its effective fee rate. Not fee divided by size. If it has unconfirmed parents, miners score the package.
- Find what the next block is clearing at. This changes hour to hour.
- Work out the shortfall in satoshis. That is what closing the gap actually costs on chain.
- Compare that against any quote. Now the decision is arithmetic rather than anxiety.
Our free checker does all five and shows the shortfall next to what the largest paid service would quote for the same transaction. We built it that way deliberately, because the comparison is the part everyone else leaves out.
The bottom line
Paid pool acceleration is a real service that does a real thing. The pools exist, the arrangement exists, we verified the hashrate ourselves at 83.4 percent, and transactions do get mined. Anyone calling the whole category fake is wrong.
But the market is tiny, about 29 accelerations a day through the dominant provider, and shrinking. Prices run 3 to 17 times the actual shortfall. The median buyer in our sample was already priced for the next block and paid anyway. And the fee environment that made acceleration valuable has not existed for over a year.
If you can raise your own fee, raise your own fee. It costs a few hundred satoshis and it works on every pool, not just the ones somebody has a deal with. If you cannot, and the delay genuinely hurts, acceleration is a reasonable purchase, and you should know what the shortfall was before you agree to the price.
Check your transaction free
We push it to twenty independent nodes, show every answer, and tell you honestly whether you need to pay anyone. Most days the answer is no.
Frequently asked questions
Do free Bitcoin transaction accelerators actually work?
Only in two situations. A free accelerator rebroadcasts your transaction to well connected nodes. If those nodes already hold it, which is usually the case, resending it changes nothing at all. Rebroadcasting genuinely helps when the transaction never propagated from your wallet in the first place, or when it was dropped from mempools after the 14 day expiry.
Do paid Bitcoin accelerators actually work?
Yes, the mechanism is real. You pay a service, the service asks partner mining pools to treat your transaction as if it paid a higher fee, and those pools include it in the blocks they find. We verified that the partner pools of the largest accelerator found over 83 percent of blocks in the last month. The transaction itself is never modified. The catch is price, not capability.
How much do Bitcoin accelerators cost?
We queried the largest accelerator's public pricing endpoint for real unconfirmed transactions. On one the genuine fee shortfall was 1,729 satoshis against a cheapest option of 10,000 satoshis, which is 5.8 times. On another the shortfall was 5,672 against a cheapest option of 17,500, which is 3.1 times. The most expensive options ran 10 to 17 times the shortfall.
Is a Bitcoin accelerator worth it?
Usually not. In the acceleration records we sample daily, the median accelerated transaction was already paying above the rate needed for the next block, meaning those buyers paid for a queue position they already held. Acceleration is worth paying for when you cannot raise the fee yourself, which mainly means an exchange withdrawal, and the payment genuinely cannot wait.
Why do Bitcoin developers dislike transaction accelerators?
Because paying miners outside the transaction weakens censorship resistance. Bitcoin Optech notes that miners with more hashrate produce blocks more consistently, so users wanting fast confirmation will pay the largest miners rather than small ones. Routine out of band payment pushes fee revenue toward the biggest pools and encourages mining centralisation.
Can an accelerator confirm a transaction that is not in the mempool?
Only if it has the full raw transaction, not just the transaction ID. If your transaction was dropped after the 14 day expiry, no node holds it and there is nothing for a pool to include. You need to rebroadcast the signed raw transaction hex so it re-enters the mempool. That is the one case where a rebroadcast service is genuinely the correct tool.