The tools that crack wallet passwords — explained plainly.

The three tools, in one sentence each

• btcrecover — purpose-built for Bitcoin-style wallets (wallet.dat, Electrum, and many more). Its superpower is token lists: you give it pieces (“name”, “year”, “symbol”) and rules, and it composes millions of likely passwords from them.
• hashcat — the general-purpose GPU cracker. Handles many common crypto-wallet formats too (via the modes below), plus masks, rules and dictionary attacks at enormous speed.
• John the Ripper — the classic open-source password cracker with excellent wordlist-mangling rules. Many wallet formats are CPU-oriented here, while some have GPU/OpenCL support depending on the format and the Jumbo build. Very accessible; a good first tool to learn on. Its full walkthrough covers the Monero .keys file, where John is the specialist tool — usually a CPU-style route, not a normal hashcat GPU workflow.

Which tool should you use?

Short version: start with btcrecover — it reads wallet files directly (including brain-wallet passphrases), and its token lists are the most effective approach for human-made passwords. Switch to hashcat when you remember the shape of the password (masks) or want dictionary + rule attacks at maximum GPU speed — the one extra step is converting your wallet to a hash line. Reach for John the Ripper when the wallet is a Monero .keys file, or when you want the clearest way to learn wordlists and rules on CPU. Whichever tool you pick, the candidate list decides the case — the hint-list guide includes a practice demo and a downloadable offline builder for all three tools. When a tool misbehaves, the troubleshooting index has the fixes, and the benchmarks page shows whether your speed is normal. And if someone offers to run this online for you? It cannot work that way.

All three are free and run offline. Follow the walkthrough for the tool you chose: btcrecover step by step, hashcat step by step, or John the Ripper step by step (Monero & keystores).

Hashcat modes for crypto wallets

These numbers come from the official hashcat example hashes list — they are the standard, publicly documented modes. They matter because they tell you exactly which format your file is and what its attack profile looks like (see how wallet encryption works).

The three attack strategies

1. 1

   Token list (btcrecover’s speciality)

   You supply fragments — names, dates, numbers, symbols you remember using — plus rules for how you combine them. The tool builds candidates like Name2020!, Name2021!, Name@2020… This is overwhelmingly the most effective approach for human-made passwords. Build your tokens with how your memory made it.

2. 2

   Mask attack (hashcat)

   When you remember the shape of the password — e.g. a capital letter, then 4 lowercase, then 2 digits — a mask like ?u?l?l?l?l?d?d limits the search to exactly that shape, shrinking the work by orders of magnitude.

3. 3

   Dictionary + rules (hashcat & John the Ripper)

   Feed a wordlist of likely words and let rules append digits, symbols, years, and case variations. Weak against truly random passwords; effective for the patterned ones.

Realistic expectations (speeds and time)

On a modern GPU, order-of-magnitude rates reach up to ~12,000 guesses per second on a low-round wallet.dat; a high-round one (the reference wallet measures ~267k iterations) runs in the low hundreds per second on CPU but can reach several thousand per second on GPU depending on the wallet parameters and tool path; a scrypt keystore sits around a hundred. That single factor decides whether “let the tool run overnight” is a plan or a fantasy. The feasibility guide walks through exactly this math.

Need more power? Rent a GPU, don’t buy one

If the math says “weeks on my own machine”, the first answer is a better clue set — but a real search sometimes genuinely needs GPU time. Renting is the standard solution: cloud GPU providers charge by the hour, and a high-end card costs far less to rent for a week than to buy.

• Rough cost picture. High-end consumer cards (RTX 4090-class) rent for roughly $0.20–$0.50/hour on GPU marketplaces like Vast.ai, Lambda and RunPod. A week of continuous search adds up fast — so check the feasibility math before renting anything.
• Estimate first. Run a short timed test or hashcat’s benchmark on the rented card before committing to a long run — the speed table in how long to crack shows what each format can actually do per hour.
• Upload only the hash line — never the wallet file. A rented machine is shared infrastructure, so your wallet file, keystore and seed stay offline on your own machine; only the extracted one-line hash (wallet2john, bitcoin2john or the hashcat-format line) and the candidate list ever leave it. Remote-compute safety is format- and extractor-dependent: some extracts are minimized for cloud testing, while others — such as Ethereum keystore extracts from ethereum2john.py — contain sensitive encrypted wallet material and the tool itself warns against sharing them. An untrusted rented machine may also learn the recovered plaintext password. So know exactly what your extract contains before sending it; when in doubt, keep it local — the warning above applies the same on rented hardware.
• Treat the rented machine as untrusted. Use a fresh SSH key, install nothing personal on it, keep it isolated from your wallet machine, and delete the hash and candidate files before you terminate the instance the moment the search ends.

Safety checklist if you try it yourself

• Work on a copy. Never run tools on your only copy of the wallet file.
• Stay offline or air-gapped. Tools don’t need the internet; an infected tool or a malicious “helpful” script does. Use a clean machine.
• Download from official repositories only (btcrecover’s GitHub, hashcat.net, Openwall for John). Third-party “crackers” and tutorials shipping binaries are a classic malware vector.
• Start small. Validate your token list on a test password first so you know your setup works.
• Watch the clock. If the math says weeks, the real answer is to let the feasibility numbers decide and go back to the search checklist — a fresh clue beats more GPU time, and the wallet files stay safe meanwhile.

Next steps

• Want to judge your case first? Is recovery technically possible?
• Need the clues to feed a token list? How your memory made it.
• Protecting the file while you work? Protect your wallet files.

Ready to actually run them? The step-by-step guides

This page explains the concepts. When you’re ready to run the tools, the walkthroughs below take you command by command — with the exact token syntax, the mode numbers and the honest limits.

Sources & references

1. 3rdIteration/btcrecover — maintained fork (source code)
2. hashcat — example hashes (wallet formats & modes)