Which Bitcoin Fee Bumping Method Should I Use?

Choosing the Right Fee Bumping Method

A Bitcoin transaction stuck in the mempool is not lost, but it will not confirm until miners find it profitable enough to include in a block. When your original fee rate falls below what the market demands, fee bumping lets you increase the effective fee so miners pick it up. Three approaches exist: Replace-by-Fee (RBF), Child Pays for Parent (CPFP), and third-party transaction accelerators. Each works differently, costs differently, and is available to different parties.

The right method depends on who you are (sender or recipient), what your wallet supports, how urgent the confirmation is, and how much extra you are willing to pay. The decision tree below covers every common scenario.

Method Comparison at a Glance

Decision Flowchart

Follow this sequence to pick the best method for your situation:

1. Are you the sender of the stuck transaction? If yes, check whether your wallet supports RBF. Most modern wallets do: Bitcoin Core, Sparrow, Electrum, BlueWallet, and Trezor Suite all support RBF. If your wallet supports it, use RBF.
2. Are you the sender but your wallet does not support RBF? Use CPFP by spending your own change output from the stuck transaction with a high fee rate. Electrum and Sparrow both offer this as a menu option.
3. Are you the recipient waiting for an incoming payment? You cannot use RBF because you did not create the original transaction. Use CPFP: spend the unconfirmed output you received with a fee high enough to cover both transactions.
4. Is the transaction extremely old, dropped from most mempools, or neither RBF nor CPFP is available? Consider a third-party accelerator as a last resort, or simply wait for the transaction to expire (default mempool expiry is 336 hours / 14 days).

Replace-by-Fee (RBF) in Detail

RBF works by broadcasting a new version of the same transaction that spends at least one identical input but pays a higher total fee. Nodes that support RBF drop the original from their mempool and relay the replacement instead.

Historically, BIP 125 required the sender to signal replaceability by setting a specific nSequence value. Since Bitcoin Core 28.0 (October 2024), full RBF is enabled by default, meaning any unconfirmed transaction can be replaced regardless of signaling. Bitcoin Core 29.0 (April 2025) removed the -mempoolfullrbf option entirely, making full RBF unconditional.

RBF is the most fee-efficient method because the replacement transaction occupies the same block space as the original. You pay only the incremental fee increase, with no additional transaction weight. The replacement must pay a fee at least as high as the original plus a minimum relay fee increment.

Note: RBF changes the transaction ID (txid). If the recipient is watching for a specific txid, they will need to be notified of the replacement. Protocols and payment processors that track txids should handle this gracefully.

Wallet RBF Support

The following table shows RBF support across popular Bitcoin wallets. With full RBF now unconditional in Bitcoin Core, any wallet that can construct a replacement transaction can use RBF, but built-in UI support varies.

For a detailed comparison of fee bumping tools across wallets, see the Bitcoin fee bumping tool comparison.

Child Pays for Parent (CPFP) in Detail

CPFP relies on miners evaluating transactions as packages rather than individually. When a child transaction spends an unconfirmed parent output and pays a high fee, miners calculate the combined (ancestor) fee rate of both transactions. If the combined rate clears the threshold for inclusion, both the parent and child confirm together.

The key advantage of CPFP is that the recipient of a stuck payment can use it. If someone sent you Bitcoin with a low fee, you can spend that unconfirmed output in a new transaction with a fee rate high enough to cover both. The downside: the child transaction adds its own weight (at minimum ~141 vB for a simple one-input, one-output child), so the total cost is higher than RBF for the same effective fee rate increase.

Recent protocol improvements have made CPFP more reliable. TRUC (v3) transactions limit the child to 1,000 vB, preventing transaction pinning attacks that could previously block fee bumping in multi-party protocols like Lightning. Ephemeral anchors extend this further by allowing a zero-fee parent to relay in a package with a fee-paying child, enabling anyone (not just output owners) to bump the fee.

Third-Party Accelerators

Transaction accelerators are out-of-band services where you submit a txid and a mining pool agrees to prioritize it. They work outside the Bitcoin protocol: the pool simply includes your transaction when it finds the next block.

The main limitation is that accelerators only work if the cooperating pool actually finds a block soon. A pool with 5% of the network hashrate finds roughly one block every 200 minutes on average, so there is no timing guarantee. During severe congestion, accelerators may not help at all.

Services like ViaBTC, BitAccelerate, and mempool.space offer acceleration, with prices ranging from free (limited availability) to $50 or more depending on urgency and transaction size. Free tiers are unreliable: ViaBTC has progressively reduced its free quota, and BitAccelerate's free tier only rebroadcasts the transaction to around 20 nodes (useful only if propagation was the problem, not fee rate). For a breakdown of services and pricing, see the transaction accelerator comparison.

Privacy consideration: Submitting your txid to a third-party accelerator reveals your transaction to that service, linking your IP address to the transaction. This is a meaningful privacy tradeoff compared to RBF or CPFP, which operate within the peer-to-peer network.

Cost Comparison

The cost of fee bumping depends on the method, the size of the original transaction, and how far the fee rate needs to increase. The following example assumes a typical 1-input, 2-output transaction (~141 vB) stuck at 5 sat/vB that needs to reach 20 sat/vB.

In this example, CPFP costs roughly 2.3x more than RBF. The gap widens with larger original transactions because the child must compensate for more parent weight. This is why RBF is always the preferred option when available.

When Each Method Fails

No fee bumping method is guaranteed to work in every situation:

• RBF fails if you no longer have access to the sending wallet, if the wallet does not support constructing a replacement, or if custom scripts prevent re-signing.
• CPFP fails if the stuck transaction has no spendable output you control (for example, if all outputs go to addresses controlled by others and you are the sender with no change output). Descendant limits in the mempool can also prevent CPFP if the transaction already has too many unconfirmed descendants.
• Accelerators fail during extreme congestion when even pool-included transactions get outbid, or when the cooperating pool does not find a block in a reasonable time frame.

If all methods fail, the transaction will eventually expire from node mempools. The default expiry in Bitcoin Core is 336 hours (14 days), after which the UTXOs become spendable again in a new transaction. Note that there is no consensus-level expiry: a transaction remains valid indefinitely if it stays in at least one node's mempool.

Avoiding Stuck Transactions Entirely

The best fee bump is the one you never need. Accurate fee estimation at the time of broadcast reduces the chance of a stuck transaction. Tools like mempool.space provide real-time fee rate data broken down by confirmation target.

For use cases where confirmation delays are unacceptable, layer-2 protocols eliminate the problem structurally. Spark transactions settle with instant finality on Bitcoin's layer 2, with no mempool, no fee market bidding, and no possibility of a stuck transaction. For payments, point-of-sale, and time-sensitive transfers, moving to a layer 2 with deterministic finality removes fee bumping from the equation entirely.

For a deeper technical dive into how RBF and CPFP work under the hood, see our complete guide to bumping stuck Bitcoin transactions.

Frequently Asked Questions

Can I use RBF if I am the recipient of a stuck Bitcoin transaction?

No. RBF requires re-signing the original transaction inputs, which only the sender can do. As the recipient, your option is CPFP: spend the unconfirmed output you received in a new, high-fee child transaction. Miners will evaluate both as a package and confirm them together if the combined fee rate is high enough.

Does RBF work on all unconfirmed Bitcoin transactions now?

Yes, on nodes running Bitcoin Core 28.0 or later. Full RBF became the default in Bitcoin Core 28.0 (October 2024), and the option to disable it was removed in Core 29.0 (April 2025). Any unconfirmed transaction can be replaced regardless of whether it originally signaled BIP 125 replaceability. However, the replacement still must pay a strictly higher fee than the original.

How much does CPFP cost compared to RBF?

CPFP is typically more expensive because it adds a second transaction to the blockchain. The child transaction consumes at least ~141 vB of block space and must pay enough to bring the combined ancestor fee rate above the confirmation threshold. In practice, CPFP costs roughly 1.5x to 2.5x more than RBF for the same fee rate target, with the gap widening as the parent transaction is larger.

Are Bitcoin transaction accelerators safe to use?

Legitimate accelerators from established mining pools (like ViaBTC) do work, but with caveats. They only help if the cooperating pool finds a block, there is no guarantee of timing, free tiers are unreliable, and you expose your transaction to a third party. Avoid services that ask for your private keys or seed phrase: a real accelerator only needs a txid.

What happens if my stuck transaction expires from the mempool?

Bitcoin Core nodes drop unconfirmed transactions from their mempool after 336 hours (14 days) by default. Once dropped, the original UTXOs become available to spend in a new transaction. The transaction is not "reversed" in a consensus sense: it simply never confirmed. If the transaction remains in at least one node's mempool and eventually gets mined, it can still confirm after the default expiry period.

Can I cancel a Bitcoin transaction using RBF?

Yes. You can create a replacement transaction that sends all funds back to your own address with a higher fee. Wallets like Sparrow and Electrum offer a "Cancel Transaction" option that does exactly this. Since full RBF is now unconditional, this works on any unconfirmed transaction, not just those that were flagged as replaceable.

How do v3 (TRUC) transactions improve fee bumping?

TRUC (Topologically Restricted Until Confirmation) transactions, also known as v3 transactions, limit the descendant tree to a single child of at most 1,000 vB. This prevents pinning attacks where an adversary attaches a large, low-fee descendant to block legitimate fee bumps. TRUC transactions make CPFP more reliable for protocols like Lightning that depend on timely confirmation of commitment transactions.

This tool is for informational purposes only and does not constitute financial advice. Data is approximate and based on publicly available information. Always verify current fee rates and wallet capabilities before attempting to bump a transaction.