Missing Order in Known Mnemonic and Address Need Assistance

Hey everyone, I got this project from a buddy of mine. He has a client with 2500 btc, and they have the 12 mnemonic words. But when they try to restore the wallet on the blockchain, it throws an error about an invalid bip39 checksum. I’m looking for a GPU-based mnemonic recovery tool because they want this project to stay confidential, so I can’t share the mnemonic words. However, I can give out the btc address. BTC address: bc1qlmal276kkvrkn36m33xvlylfgqspcdkp0l2zsz I’m really needing help to recover this. I’ve been trying to check permutations for the lost word up to the first thousand addresses, but a GPU program would be way faster since CPU programs are just taking ages. I tried to set up jhoncartnell97's bip39-solver-gpu, but I ran into some build errors. For anyone who can help with recovering the mnemonics and btc, a nice reward is on the table. Thanks a lot! Kosal Ram ICONBTCX

Yes it will actually give an invalid checksum because the words are arranged in wrong order. You can use the btcrecover to get the arrangement check here https://btcrecover.readthedocs.io/en/latest/. With the address format know it will be a bit easier, what you do is follow the command line for a segwit address (P2PWKH) recovery because that’s the format for your address

Hai guys,            Thanks for your suggestions but btcrecover is replacing seeds with unknown words like 4 phases i tried its not that much worth the shot i think and finder outer also bit slow and systems hang up, its also not that much effective because we dont know exact path location of address means m/84/0/0/0-1000 it may be any where in the first 1000 address,  i tried all these methods please i need if someone able to developed @jhoncartnell97 bip39-solver-gpu rust code when i am trying to build programn it throwing erros. Guys please dont worry about reward i am a man of my word sure whom help i will definitely give some good reward. Thank you U Kosal Ram ICONBTCX

This can be done with itertools in Python. I have already mixing script to do that.

Mizogg bro has given me one its excellent one but taking long time it takes 2 days just to move to fourth position he developed an excellent code permutation and iter based We need some solid gpu code bro we have lot of gpus for scrambled mnemonic recovery

Does the client know the position of any of the words of the seed phrase or does he not have any idea what goes where? It would certainly speed things up if he knew, at least, some of the words. It's good that you have the funded address. I still think that BTCrecover is the way to go here. You can take a look at this guide that concerns the recovery of seed words. If nothing else works, you could use a trusted paid service. Dave's wallet recovery services have been around for years, and he knows what he is doing. If you and your friend can't manage yourselves, contact Dave.   https://bitcointalk.org/index.php?topic=240779.0 https://www.walletrecoveryservices.com/ Someone who knows much more than me about these things estimated a long time ago that it would take 200 years to bruteforce the order of a 24-word seed. Yours has 12 words, but it would still be a difficult task. And there have been advancements in computer tech since 2017. Anyways, good luck!

With the premise "you have enough computing power to try 100 trillion combinations per second" taken from your example it would be an easy task as 12 words allow only  4.790016∗108 distinguishable permutations rather than 6.2044840173323943936∗1023 as in the  case of 24 words. What is the realistic computational power to perform   a single  12 of 12 permutation?

Did you take a look at https://btcrecover.readthedocs.io/en/latest/ and https://btcrecover.readthedocs.io/en/latest/Usage_Examples/2020-05-02_Descrambling_a_12_word_seed/Example_Descrambling_a_12_word_seed/ Should give you a general idea of what to do, how to do it and how long it will take. With videos. -Dave

Hai guys thanks for your support As seeing the calculations you had shared with me I would like to mention the words are not 24 its 12 all we need to do is permutation check in those words each words should replace only once in the sequence so 12p12 is just 479001600 combos I think its not impossible. THanks for your concern all i need is a gpu based permutation check

Thanks. I watched that "Recover Scrambled 12 Word Seeds with BTCRecover" video and and was convinced in my correctness - the set  12 known words can be unscrabled with virtually no effort even with use of machine of moderate  computational power. It took something around 50 minutes for guy to descramble BIP39 Seed phrase and even less time in the case of Electrum seed.

I am surprised it's that quick. But on the other hand, all words are known, and there is no need to bruteforce unknown words in any positions. Is it a requirement for such a quick recovery to know an address that belongs to the seed or the script type?

In fact , this is not surprising in those   4.790016∗108 chances. Address that belongs to the seed is among those  parameters taken required by seedrecover.py . Amidst them are also address limit  and derivation path. The latter is as important (I think) as address itself.

The post above this quoted reply supports select GPUs in the form of --enable-opencl command line arg. But it is still experimental so it may or may not work; in my test, it worked in electrum wallet password recovery but not in BIP39 seed. You likely have a different GPU and specs anyways; in any case, any decent CPU can still solve it in reasonable time. Additionally, since you don't know the derivation path of the address, using a larger --addr-limit can drastically reduce its speed. Just find a way to reduce the addresses to cross-check by utilizing options like --mpk to significantly speed it up if you know your xpub and even better with --bip32-path.

Total number of permutations: 479001600 That's 479 M.

OP needs to find SEED relevant to address that holds his stash rather than to get the full  list of all possible  permutations pertaining to his 12 scrambled words , thus the script you suggested is, in fact, useless for him. To find what OP wants he needs to run seedrecover.py (or similar script) that perfectly does his job.

This is trivial to do using btcrecover. First of all create a new text file with one word of the seed phrase per line. Something along the lines of this: Then run the following command using btcrecover: This will check every address between m/84/0'/0'/0/0 and m/84'/0'/0'/0/999 for every valid combination of your 12 words. If this is very slow on your hardware, you can experiment using GPU acceleration by following the instructions here and adding the relevant arguments to your command: https://btcrecover.readthedocs.io/en/latest/GPU_Acceleration/

Thanks brothers for your valuable support, I counting on btcrecover seedrecover if i want to search of the address in definitive path shall i have to give option --bip32-path m/84'/0'/0'/0 or shall i avoid this, and please guide me how to enable multiple gpus using --opencl-devices ID1 ID2 like that with an example i have four gpus THanks for your valuable support and hope its good and we can able to recover i promise everyone who tried to help surely get rewards. Fingers crossed brothers https://imgur.com/a/NWBT4Jh why its not working [moderator's note: consecutive posts merged]

Does it run if you leave out the --enable-opencl argument? If it does, then it probably means you haven't installed the correct drivers or libraries for your GPU.

It'll be faster with it, in general it searches for the paths listed in "/derivationpath-lists/BTC.txt" regardless of the address type, with --bip32-path arg or by commenting-out the unneeded lines, it will only search those 1000 address_index in one derivation path. But you'll have to make sure that your wallet is using the specified path or you'll get false-negative result.

Thanks guys for your support, Even though i have drivers and everything updated but its showing 0 p/s but i can hear gpus started whirling shall i have to wait. Thank you