BIP39 Passphrase Recovery (Or Hidden Wallet Password, 25th Word) For Ledger, Trezor, Keepkey, etc
If you have ever entered your recovery phrase into a wallet and seen the wrong addresses appear, you may not have made a mistake in the seed. You may have made a mistake in the passphrase.
That is the problem I cover in this article: recovering a lost or mistyped BIP39 passphrase, also called a hidden wallet password or “25th word,” when your 24-word seed is correct. Watch the video if you want to see the full walkthrough, but this written version gives you the same practical process in a form you can come back to later.
This is especially relevant for Ledger, Trezor, Keepkey, Electrum and similar wallets that let you add an optional passphrase to your BIP39 seed phrase. A passphrase is useful, but it is also one of the easiest things to mistype or forget. If the seed is correct but the wallet does not show the addresses where your funds are, a passphrase error is worth checking before you start blaming the seed itself.
Stuck with a wallet recovery?
I maintain BTCRecover and have helped people recover seed phrases, passphrases and wallet passwords since 2019. Book a one-hour video session and I’ll walk you through doing the recovery yourself, without sharing your seed. Or, if you’d rather I run it, a trusted recovery is charged only if it succeeds.
Private sessions: US$100 per hour. Trusted recovery: 10% of recovered funds. I never contact people first and will never ask for your seed phrase over chat or email.
What You Need Before Starting
Before you try this, there are two important requirements.
First, you need a correct BIP39 seed phrase. This method is not for recovering a damaged or mistyped 24-word seed. It assumes your seed words are correct and that the only missing piece is the passphrase.
Second, you need at least one known public address from the wallet account you want to recover. The tool works by testing possible passphrases against known addresses. In other words, it tries different passphrase combinations and checks whether any of them produce an address you already know belongs to your funds.
A simple example: suppose you bought Ethereum on Coinbase and remember sending it to a specific wallet address. That address becomes the target. If your seed is correct but you cannot find that address because of a passphrase mistake, this process can help identify what the passphrase actually was.
You should also be clear about which account you are trying to recover. Wallets such as Ledger Live use different derivation paths for different accounts. If you added multiple accounts in Ledger Live or another wallet, each one may use a different path. The tool needs the correct derivation path for the account you are searching.
Why Passphrase Mistakes Happen
A passphrase is an extra word or phrase that changes the keys generated from your seed. It can add protection, but it also adds one more thing to remember and type correctly.
This is why Ledger and Trezor treat passphrases as an advanced feature. If you rely on memory instead of writing it down, a small mistake can create a completely different wallet. You may enter the correct 24-word seed, then mistype the passphrase, and end up looking at addresses that have no transactions in them.
Some wallets make this even easier to mess up. For example, Electrum does not always ask you to type the passphrase twice for confirmation. If you are in a hurry, it is possible to enter a typo without noticing it immediately. That mistake can stay hidden for months if you do not carefully validate the addresses before moving funds.
The good news is that if your seed is correct and you know at least one public address, recovering the passphrase is often possible using BTCRecover.
Setup: Use an Air-Gapped Environment or Windows
For this process, I recommend using the SegWit fork of BTCRecover. You will need to set up a safe environment first.
The safest option is an air-gapped Ubuntu 18.04 machine. That means a computer that is not connected to the internet while you run recovery tools and enter your seed phrase. I cover the setup in another video, but the key point is this: do not type your real seed into an online tool or a compromised system.
If you are more comfortable with Windows, you can also run some of these tools directly on Windows. It may be less ideal from a security perspective, but it is better than doing nothing if that is what you are able to manage. The important thing is to keep the process isolated and avoid exposing your seed to unnecessary risk.
Before using real data, I strongly suggest testing the tool with example files first. This helps you confirm that the syntax works, that you understand the options, and that you know what a successful result looks like before putting in your own information.
Password Lists vs Token Lists
Once BTCRecover is running, there are two main ways to describe possible passphrases:
- Password list
- Token list
I recommend starting with a password list because it is easier to understand and easier to control.
Password List
A password list is simply a file containing lines of possible passphrases. Each line is one candidate passphrase that the tool will test.
This works well when you have a reasonable idea of what the passphrase was, but there may be a typo or small variation. For example, you might remember the general structure but not be sure about spelling, capitalization, year, symbol placement or word order.
You should include likely variations yourself. The tool will not automatically invent every possible misspelling for you unless you configure it to do so. So if you think you may have typed a common misspelling, included the wrong year, capitalized one letter differently, or swapped two letters, add those versions to the list.
The big advantage of a password list is that it scales in a predictable way. If you double the number of lines in the file, the run time roughly doubles. That makes it easier to estimate how long a search will take and less likely to accidentally create a search space so large that it runs for weeks or years.
For many people, this is the best first step. Even a fairly long password list can be practical if you are narrowing down realistic possibilities.
Token List
A token list is more advanced. Instead of listing full passphrases, you define smaller pieces, or tokens, that might make up the passphrase. For example, you might remember certain words, numbers, symbols or patterns that were part of the password, but not the exact order or length.
Token lists can be powerful because they can model complex passwords built from repeated building blocks. However, they are also much easier to misconfigure. If the token list is too broad, BTCRecover may try an enormous number of combinations. In some cases, that could take years to complete.
To make a token list usable, you usually need to add constraints. For example:
- The passphrase likely started with one of a few specific words.
- It ended with a year or another known block.
- A symbol appeared somewhere in the middle.
- Certain tokens were not allowed at certain positions.
- The total length was within a certain range.
These constraints reduce the number of possible combinations and make the search manageable. But token files can become complicated quickly. If you are new to BTCRecover, I would not start here unless a password list has already failed or you have a very strong reason to use tokens.
Suggested Safe Workflow
Here is the workflow I recommend:
-
Prepare an isolated environment.
Use an air-gapped Ubuntu machine if possible. If you must use Windows, be extra careful and avoid online tools with your seed. -
Download the correct BTCRecover repository.
For this article, we are using the SegWit fork of BTCRecover. -
Test the tool first.
Run it with example files before entering your real seed phrase or target address. This helps you catch syntax errors early. -
Start with a password list.
Put likely passphrases and realistic typo variants into a simple text file, one candidate per line. -
Use typo options carefully.
Enable common typo checks such as caps lock mistakes, swapped letters, repeated letters, deleted letters, case changes and nearby-key errors. -
Set the correct wallet type and derivation path.
For example, Ethereum uses a different path than Bitcoin. If you are searching a specific account in Ledger Live or another multi-account wallet, make sure the path matches that account. -
Use a sensible address limit.
Start with one address if you know exactly which address to find. You can increase it later if needed. -
Only move to token lists if necessary.
If the password list does not succeed and your memory is too vague for full candidates, then build a constrained token list.
This order matters because it reduces the chance of wasting hours on a badly configured search.
Example 1: First Address and Short Password List
In the first example, we are looking for the first address in the first account on the default derivation path. The seed is known to be correct. We have one target public address, and we have a small password list containing our best guess plus likely typo variations.
The command uses several typo options:
- Check for caps lock mistakes.
- Check for swapped letters.
- Check for repeated letters.
- Check for deleted letters.
- Check for capitalization changes.
- Allow up to three typos.
- Use a typo map for nearby keyboard keys.
We also set the wallet type to Ethereum and limit address generation to one, because we are only looking at the first address in this example.

When run, BTCRecover asks for the extended public key if available. If you do not have an xpub file, it can work from a known address instead. In this example, we provide the target address and the correct seed phrase.
The tool then tests the candidates against the expected address. In this case, it found the passphrase in about 30 seconds. The important point is not the specific password itself, but that a short list plus good typo options can quickly recover a simple mistake.
Example 2: Longer Password List
In the second example, we use a different passphrase and therefore a different set of generated addresses. We are looking for another known public address. This time, our memory is less precise.
We think the passphrase may have included a common word related to Ethereum, but we are not sure about spelling or capitalization. We also are not certain which year was used. So instead of guessing one exact phrase, we create a longer password list with multiple realistic variants:
- Correct spelling and wrong spelling.
- Different years.
- Capitalized and non-capitalized versions.
- Variants that might have been typed in a hurry.
- Possible related terms if the original word was not exactly what you remember.
The command is similar to the first example, but it uses this larger password file. Because there are more candidates, it takes longer. In the video, this run took about 11 minutes and found the passphrase before reaching the end of the list. It could have taken much longer if the correct candidate had been near the bottom.
This is a good example of why password lists are practical. Even when you do not remember the exact passphrase, listing realistic possibilities can still be manageable, especially compared with an unconstrained token search.
Example 3: Token List, Tenth Address, Second Account
The third example shows how to use a token list in a more advanced scenario. This time we are not looking for the first address. We are looking for the tenth address, which is common for wallets that generate new receiving addresses over time. Bitcoin users often see this behavior because many wallets give you a new address each time you request one.
For Ethereum, it may mean scrolling through several pages of addresses in a wallet such as MyEtherWallet and selecting a later address. You might have used the first few addresses for different things and only remembered a later one.
We are also using a different derivation path because this example represents a second account. If you use Ledger Live or another multi-account wallet, adding a new account changes the derivation path. That matters here because BTCRecover uses the derivation path to generate the correct sequence of addresses.

In this example, the command uses:
- A token file instead of a simple password file.
- An address generation limit of 20.
- A specific derivation path for the account being searched.
- The same general wallet type and typo options as before.
One important point is that increasing the address limit does not increase run time in a strictly linear way. Going from checking one address to twenty addresses will not usually make the search twenty times slower. In the video, doubling the number of generated addresses increased processing time by only about 20%. That means it is often practical to check more addresses if you are unsure which one was used.
For Bitcoin especially, where receiving addresses may change frequently, setting the address limit to 10, 20 or even 50 can be a sensible safety measure without making the run unreasonably long.

In this example, I initially made a typo in the derivation path. That caused the search to run for about two hours before I realized the mistake. After correcting the path and rerunning it, the tool found the passphrase in about 10 minutes. A full correct run would have taken around three hours, but finding it early saved time.
This is a very practical lesson: if your derivation path is wrong, you can waste hours searching for an address that cannot possibly appear on that path. Double-check the account and path before starting a long search.
Derivation Path Typos Are Expensive
The most common “silent failure” in this kind of recovery is not necessarily a bad password list. It is using the wrong derivation path.
If you are searching for an address from your first account, but the tool is configured for the second account, it may run through thousands or millions of passphrase candidates and find nothing. That does not mean the passphrase is unrecoverable. It may simply mean the search was looking in the wrong place.
Before running a long job, confirm:
- Which wallet type you are using.
- Whether you are searching Bitcoin, Ethereum or another coin.
- Which account number you need.
- Whether the address belongs to the default path or a custom path.
- Whether you are checking receiving addresses, change addresses or both, if relevant.
- Whether your known address is from the same wallet and account you think it is.
A small typo in the path can turn a 10-minute search into a multi-hour dead end.
Security Reminder: A Passphrase Is Not Magic
A BIP39 passphrase adds protection, but it should not be treated as an invincible layer of security on its own. If your 24-word seed is compromised, a short or guessable passphrase may not save you. The same goes for common phrases, simple words, predictable patterns or passphrases that are easy to brute force.
The safest approach is still:
- Keep the 24-word seed offline and secure.
- Use a strong, unique passphrase if you choose to use one.
- Write down the passphrase in a safe place if your wallet supports it.
- Validate addresses before sending funds.
- Test recovery on a small amount first when possible.
- Avoid online “seed checkers” or unknown recovery websites.
If you are dealing with lost funds and want guided help, I also offer recovery services and consultations for people who need practical support with wallet recovery issues. But if your seed is correct and you know a public address, the process above can often be done yourself using BTCRecover in a safe environment.
Conclusion
If your 24-word BIP39 seed is correct but your wallet shows the wrong addresses, do not immediately assume the seed is ruined. A passphrase mistake is common, especially with wallets that allow an optional hidden password or 25th word.
To recover it:
- Confirm your seed phrase is correct.
- Identify at least one known public address from the target account.
- Use an isolated environment such as air-gapped Ubuntu if possible.
- Start with a simple password list containing realistic candidates and typo variants.
- Enable sensible typo options.
- Set the correct wallet type, derivation path and address limit.
- Only move to token lists when you need more complex modeling and can add strong constraints.
Password lists are easier to understand and scale predictably. Token lists are powerful but require careful limits. And no matter which method you use, getting the derivation path right is essential. A small typo there can waste hours of processing time without finding anything.
If this helped, leave a reply with your questions or results. If you want to see the full screen-by-screen walkthrough, watch the video linked above.
Stuck with a wallet recovery?
I maintain BTCRecover and have helped people recover seed phrases, passphrases and wallet passwords since 2019. Book a one-hour video session and I’ll walk you through doing the recovery yourself, without sharing your seed. Or, if you’d rather I run it, a trusted recovery is charged only if it succeeds.
Private sessions: US$100 per hour. Trusted recovery: 10% of recovered funds. I never contact people first and will never ask for your seed phrase over chat or email.
Found this useful?
These guides are free and ad-free. You can support them by sharing them, subscribing on YouTube, or sending a tip.







