, , , , ,

Recover Scrambled 12 Word Seeds with BTCRecover (BIP39)

If something can be brute forced in about the same length of time it would take you to have a coffee break, then you probably need a better way of securing your offline seed backup.

In this article, I’m walking through how to use BTCRecover to descramble randomized 12-word BIP39 or Electrum seeds. These are seeds that someone may have stored offline and deliberately mixed up the word order in an attempt to stop another person from using them if they found the backup.

Watch the video for the full demonstration, but this written version gives you the practical steps, command flags, token list concepts, and performance expectations in one place.

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.

Why scrambled 12-word seeds are not secure

The short answer is that a randomly scrambled 12-word seed is much weaker than people think.

If someone has all 12 words but does not know the correct order, they can use BTCRecover to try every possible permutation of those words and test them against an address or wallet type. For a full 12-word descramble where no word positions are known, this is a large search space, but it is still small enough that modern hardware can work through it in a practical amount of time.

That matters for two reasons:

  • If you have scrambled your own seed and forgotten the order, BTCRecover may be able to recover it.
  • If you are using word scrambling as your main security measure against someone finding your offline backup, you should assume that is not strong protection.

This article is useful if you have found yourself in a position where you scrambled your 12-word seed, kept the backup offline, and now cannot remember how you mixed up the words. It is also useful if you want to understand why this backup method is insecure.

Who this recovery guide is for

This guide is aimed at people who:

  • Have a 12-word Electrum or BIP39 seed phrase.
  • Know all the words but not the correct order.
  • Want to use BTCRecover to test possible word orders against an address.
  • Are trying to understand how much time and hardware is needed for different recovery scenarios.

It is especially useful if you have a partial memory of the seed, such as knowing where some words should be or what some words start with. That kind of information can dramatically reduce the search space.

If you are not comfortable running command-line tools yourself, I also offer recovery services and consultations: https://cryptoguide.tips/recovery-services-consultations/

Downloading and installing BTCRecover

The first thing you need to do is download BTCRecover from the repository I maintain. I have been maintaining it for a few months, bringing in updates and new features because the original repositories and forks are not being actively maintained.

GitHub repository page for 3rdIteration/btcrecover
The GitHub repository used for this guide

You need to get it from my repository because that is where the current instructions, examples, token lists, and templates live.

If you have never used BTCRecover before, you will also need Python 3 and a few required packages. The installation instructions are in the GitHub repository.

GitHub repository page with Clone or download button
Clone or download the repository

Once it is downloaded, you can work through the included examples. I deliberately include several different examples for different wallet types and cryptocurrencies, including:

  • Electrum Segwit
  • BIP39 wallets
  • Native SegWit Litecoin
  • Ethereum
  • Legacy BTC examples in the documentation

These examples are useful because they give you a working starting point that you can adapt to your own situation.

UsageExamples.md documentation file
Example commands and templates included in the repository

How token lists work

BTCRecover uses token lists to describe what is known and unknown about the seed.

A token list is a formatted text file that tells BTCRecover which words are fixed, which positions are unknown, and which possible words can be tried in each position. In other words, it defines the search space.

The syntax for creating and formatting token files is covered in the repository documentation. The examples included with the tool use different token lists so you can see how this works in practice.

tokenlist_file.md documentation
Documentation explaining token list files

A full 12-word descramble where no word positions are known uses a very large token list because every position is unknown. A constrained example, such as one where you know some words or their starting letters, can be much faster because the search space is smaller.

fixedwords_bip39.txt token list file
Example BIP39 fixed-word token list

Performance expectations

Performance depends heavily on your hardware and how constrained the token list is.

In the video, I mention spinning up BTCRecover on a 48-core Linode. On that kind of machine:

  • A full Electrum seed descramble can be done in under about 15 minutes.
  • A full 12-word BIP39 seed descramble should take no more than about 50 minutes, assuming the correct seed is literally the very last one tested.

On my local machine in the example shown, a full Electrum Segwit descramble took about 11.5 minutes. On the 48-core Linode, I expect that same kind of job to take around two minutes.

BTCRecover running on a 48-core Linode
Live progress output from a high-core-count server

Brute-forcing with BTCRecover is ultimately about probability. If you have a randomly scrambled 12-word seed, there is roughly a 50% chance the correct order will be found before the halfway point of the search and a 50% chance it will be found after that. The more information you have about word positions or starting letters, the faster the process becomes.

Example: Electrum Segwit descramble

The first example I show is for an Electrum Segwit wallet with a randomized 12-word seed where no correct order is known.

Electrum Segwit BTCRecover command in Windows Command Prompt
Example command for recovering a scrambled Electrum Segwit seed

The important parts of the command are:

  1. No duplicate checks
    This helps save memory by skipping some duplicate checking.

  2. Mnemonic length set to 12
    BTCRecover needs to know how many words the seed contains.

  3. Language specified
    The example uses English.

  4. Security warning disabled
    The command includes a flag to disable the security warning.

  5. Wallet type set to Electrum 2
    This matters because different wallets derive addresses differently.

  6. Address limit set to one
    In this example, the target address is the first address in the wallet. If you do not know how early an address appears in the wallet, you may need to increase this limit.

  7. Token list path specified
    This points BTCRecover to the token file describing which words are known or unknown. In the example, it uses a relative path that works if you unzip the repository as provided.

  8. Derivation path for Electrum Segwit
    Because this is an Electrum Segwit wallet, the command must use the correct derivation path for that wallet type.

Electrum 2 Segwit seed recovery command with flags
Command-line invocation for an Electrum Segwit example

When you run it, BTCRecover first counts how many possible passwords exist and then starts testing them. This counting step can take a little while.

BTCRecover startup output counting possible passwords
Startup output before the search begins

In the example shown, this particular Electrum Segwit descramble took about 11.5 minutes on my machine. On a strong multicore server, it would be much faster.

Example: Litecoin native SegWit BIP39 recovery

The second example is more constrained and probably more realistic for someone who remembers part of their seed.

Suppose you have Litecoin in a native SegWit address from a BIP39 wallet, but the 12 words are scrambled. You may know that some words belong in certain positions or that some words start with particular letters. That information can be encoded in the token list.

Litecoin native SegWit BTCRecover command
Example command for a Litecoin BIP39 native SegWit address

The command is similar to the Electrum example, but with important differences:

  • The wallet type must be set to BIP39, not Electrum.
  • The target address is a native SegWit Litecoin address.
  • The derivation path is the BIP32 path used for Litecoin.
  • The token list reflects whatever partial information you have about word positions or starting letters.

In this example, the correct seed was found in 23 seconds because the token list already narrowed things down significantly. That shows how much a small amount of remembered structure can help.

Ethereum recovery and derivation paths

Ethereum is similar to BIP39 recovery, but you must be careful with wallet type and derivation path.

For Ethereum, you need to specify the wallet type as Ethereum, not just BIP39. This is because Ethereum derives addresses differently from many Bitcoin-based wallets.

Ethereum seed-recovery command examples
BIP39 seed recovery examples for Ethereum and other chains

The default derivation path used in the example is the common one used by many modern wallets, including Trezor and Muon-style setups, and it is also what Ledger Live uses now.

However, there is a big gotcha: older Ledger devices or some third-party wallets such as Koinomi may use nonstandard Ethereum derivation paths. If your wallet does not use the path you are testing against, recovery can fail even if the seed itself is correct.

Ethereum example command highlighted in documentation
Ethereum example from the usage documentation

When running an Ethereum descramble, I also show how to skip the password counting step using --no-eta. This starts processing immediately and shows a live rate of passwords per second.

BTCRecover Ethereum command with --no-eta option
Running an Ethereum recovery with no ETA

This flag is useful in two ways:

  1. It skips the initial password-counting phase, which can be slow for large token lists.
  2. It gives you a benchmark by showing how many thousands of passwords per second your machine is testing.

If you are just descrambling a 12-word seed and already have a reasonable estimate of how long it will take, --no-eta can save time. On my local PC in the video, running through the complete address space for a full 12-word descramble takes about six hours. A high-end multicore machine such as a strong Linode or Threadripper system can do it in around 50 minutes.

Longer seeds and search-space limits

Descrambling a 12-word seed is relatively straightforward because the number of permutations, while large, is still manageable on good hardware.

Descrambling 15, 18, or 24-word seeds is an entirely different matter. The search space grows exponentially as the number of unknown words increases.

That said, partial knowledge can make longer seed recovery much more practical. For example, if you have a 24-word seed and you know the correct positions for 12 of those words, the problem becomes similar to descrambling a 12-word seed where none of the word positions are known. In other words, knowing even part of the structure can make a huge difference.

The key point is this: as the number of unknowns increases, the search space increases exponentially. If you have any partial memory of your seed—word positions, starting letters, or groups of words—you should encode that into the token list before running the recovery.

Troubleshooting and help

If you run into trouble with BTCRecover, your best options are:

  • Leave a reply on the YouTube video.
  • Open an issue in the GitHub repository.

I will do my best to help where I can. If you use this tool to recover a large amount of crypto, I also encourage you to consider sending a donation to support continued work on the channel and ongoing maintenance of tools like BTCRecover.

Conclusion

Scrambling the word order of a 12-word seed is not strong security. With the right token list, wallet type, derivation path, and hardware, BTCRecover can test possible word orders quickly enough that this method should not be relied on as your main protection against someone finding an offline backup.

That said, if you have scrambled your own 12-word Electrum or BIP39 seed and forgotten the order, BTCRecover is a practical tool for recovering it. The most important things are:

  • Use the correct wallet type.
  • Use the correct derivation path.
  • Build a token list that reflects everything you remember.
  • Set an address limit high enough if you do not know where the target address sits in the wallet.
  • Expect full 12-word descrambles to take hours on a normal PC, but much less on strong multicore hardware.

If you only have a complete 12-word set and no other information, recovery is still possible, but it will depend heavily on your machine and how far through the search space the correct order falls. If you have partial knowledge of word positions or starting letters, the process can become dramatically faster.

Questions from viewers

Common questions from the comments on this video, with my answers.

What is the realistic limit for how many words I can be missing or scrambled?

For a 12-word seed, you can realistically recover it if you are missing up to 3 words or have them fully scrambled. Missing 4 or more words makes it practically impossible with current technology. For 24-word seeds, the math becomes much harder; while you can recover with a few missing words, fully scrambling all 24 would take billions of years. The realistic limit for descrambling a longer seed is around 15 words.

Do I need my wallet address to run BTCRecover?

No, you do not strictly need the address. If you don't know your address, you can use an AddressDB (Address Database) which contains every address ever used on that blockchain. This allows the software to verify if a generated seed matches any known transaction history. This works for Bitcoin, Ethereum, and other standard BIP39 wallets.

How long will the recovery process take?

It depends heavily on your hardware and the complexity of the search. On a modern high-end CPU or GPU, a full 12-word descramble might take about an hour. On older local machines, it could take several hours or even days. If you are only missing one word, it takes significantly less time than a full scramble. Note that Electrum 1 wallets are encrypted differently and are much slower to check than standard BIP39 or Electrum 2 wallets.

Does BTCRecover work with non-Bitcoin wallets like Ethereum, Exodus, or Trust Wallet?

Yes, as long as the wallet follows standard BIP39/44 derivation paths. You just need to select the correct wallet type in the command (e.g., 'ethereum' for MetaMask, Trust Wallet, or Binance BEP20 tokens). Some specific wallets like Guarda or certain alt-coins that do not follow standards may not be supported or require custom modifications. For most mainstream wallets, it works out of the box.

Is scrambling my seed phrase a good security measure?

No, scrambling a 12-word seed offers zero extra protection. It can be brute-forced in about an hour with GPU acceleration. If you want to secure your backup against theft, you should use a BIP39 passphrase (an extra word or phrase added to the seed), not by rearranging the existing words.

What should I do if I get a 'seed not found' error?

This usually means either the seed doesn't actually match the address you provided, or there is an error in your setup. My suggestion is to generate a test seed first to validate that your process and commands are working correctly before running the real recovery. Also ensure you are using the correct derivation path and wallet type for your specific crypto.

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.

Stuck on this?

Book a one-hour session and I’ll walk you through it, or ask about a trusted recovery.

Recovery guides

More guides