Guide 08
Is wallet password recovery possible? The equation.
Recovery is arithmetic: how many passwords can your wallet type check per second, divided by how many passwords you still need to try. Here is exactly how that plays out per wallet format.
By Robbert Bink · ~7 min read · Published 6 August 2026 · Last reviewed: 14 August 2026
Usually yes — when the password was human-chosen and you have hints, recovery can take days to weeks. A genuinely random 16-character password is generally infeasible to brute-force — no service, GPU farm or “hacker” changes that. No wallet file left? The real exceptions — seed phrase, custodial resets, brain wallets — are covered on recovery without the wallet file.
How wallet encryption actually works
Wallet files don’t store your password. They store an encrypted key, and your password is run through a key-derivation function (KDF) that turns it into the decryption key. The KDF is deliberately slow — it is the wallet’s whole defence against guessing. The number of times it runs per attempt is the iteration count:
- Low iterations = fast to test = easier recovery. Older wallets are faster to attack.
- High iterations / scrypt = slow to test = harder recovery. Modern wallets are slow by design.
“Hashes per second” is the real speed. Modern GPUs try millions of guesses per second against simple hash types — but only tens to tens of thousands per second against deliberately slow wallet KDFs.
The wallet formats, side by side
For a one-glance answer on every major wallet — reset, crackable offline, or impossible — start with the master list of recoverable wallet passwords, then come back here for the speed numbers. When the verdict is feasible, the complete DIY route is how to recover a wallet password yourself.
| Wallet type | Encryption / KDF | hashcat mode | Speed class (modern GPU)* |
|---|---|---|---|
Bitcoin Core wallet.dat (old) |
iterated SHA-512, ~25k+ rounds (machine-tuned; reference 68k) | 11300 | Thousands of guesses/sec (up to ~12k on a modern GPU) |
Bitcoin Core wallet.dat (modern) |
iterated SHA-512, machine-tuned rounds (reference ~68k–267k) | 11300 | Hundreds to low thousands of guesses/sec |
| Electrum v1–4.x | Encrypted keystore — PBKDF2 (2.8+), double SHA-256 (older) | 16600 / 21700 / 21800 | Low thousands of guesses/sec |
| Ethereum keystore (scrypt) | scrypt (default N=262144, r=8, p=1) | 15700 | Around a hundred guesses/sec |
| Ethereum keystore (PBKDF2) | PBKDF2-HMAC-SHA256; iteration count stored in crypto.kdfparams.c |
15600 | Often low-thousands/sec, but estimate from the actual c value |
| BIP38 encrypted key | scrypt (N=16384, r=8, p=8) | — | Tens of guesses/sec |
| Brain wallet — SHA-256 (bitaddress.org) | SHA-256(passphrase) → key, no KDF | — | Millions of guesses/sec on GPU (dedicated tools — no hashcat mode) |
| Brain wallet — scrypt (WarpWallet/Memwallet) | scrypt + optional email salt | — | Tens of guesses/sec |
*Speed classes are order-of-magnitude estimates on a high-end consumer GPU (e.g. RTX 4090 class). Exact rates vary with hardware, driver, and hashcat/btcrecover versions. What matters for your case is the category — thousands vs hundreds vs tens of thousands.
What “impossible” actually means
Brute force scales with entropy — the number of possible passwords. Roughly:
- 8 characters, human-chosen (word + number + symbol): often recoverable in hours-to-days if you have hints.
- 10 characters, human pattern: days-to-weeks with good clues and the right masks.
- 12 characters, mixed random: centuries on consumer hardware — effectively impossible in practice.
- 16+ characters, random (password manager): beyond any realistic hardware — no tool changes this.
- Lost seed phrase (12–24 words): a seed cannot be reconstructed from nothing — find a copy via the search checklist. If the seed is intact but a passphrase (25th/13th word) was used, that IS recoverable and this site covers it.
Anyone who “guarantees” recovery of a random password is overpromising
The maths is public and it does not care about marketing. A promise to crack an unknown 16-character random password should be treated with caution — the real answer is that no tool or service can change the search space.
The four levels: possible vs feasible vs sensible vs impossible
“Can it be recovered?” is not one question — it is four, and they are often confused:
The distinction matters because most “is it recoverable?” questions collapse two different things: can a tool test this password? and will my password be found in a useful lifetime? The calculator answers the second with your numbers, and the measured speeds page shows the rates behind it — hardware, method and dates included.
The three factors that decide your case
1 · What you remember
Fragments, patterns, lengths, reused words. Every clue shrinks the search space from astronomical to manageable. This is the biggest lever you control — rebuild it with the memory method.
2 · The wallet format
Low-round wallet.dat? Thousands of guesses per second on a GPU. A modern Ethereum keystore with scrypt? Around a hundred. Format decides the speed; speed decides the odds.
3 · Randomness
Human-chosen passwords are guessable by design. Random passwords are not. No amount of hardware changes that — the line is the line.
What tools like btcrecover and hashcat do
Open-source tools such as btcrecover and hashcat automate the guessing: they read your wallet file, generate candidate passwords from masks and token files, and test them at the wallet’s KDF speed. They are free, offline, and well documented. But they are tools, not magic:
- Masks — structural templates like
?u?l?l?l?l?l?d?d?d(capital + 5 letters + 3 digits). - Token files — your candidate list from the memory guide, plus year/number/symbol mutations.
- Passphrase attack — your known words in every order, with every separator.
Success almost always comes from clues narrowing the space — never from raw brute force. That is why the diagnosis on the homepage starts with what you remember.
Running tools safely
Only ever run recovery tools on a read-only copy of your wallet file, on a machine you trust, offline. Never upload the file to a public checker or unknown website, and never hand it to anyone without a contract — protect your wallet files first. And if the password isn’t actually the problem — a dead device, a corrupted file or a missing seed — check the seven cases where cracking can’t help before you start.
Format-specific deep dives
Your wallet type changes everything. Jump to the specialist page for yours:
Frequently asked questions
Is any wallet password technically uncrackable?
Yes: genuinely random passwords with roughly 80+ bits of entropy — for example 14+ characters from a full character set — are generally infeasible for any realistic hardware, including well-funded attackers. No tool or service changes that; anyone promising otherwise is lying.
What single factor most improves my odds?
What you remember about how the password was constructed. A clue that narrows the search space by a factor of a million is worth more than any GPU — which is why memory reconstruction comes before hardware.
How long should I keep trying before concluding it is impossible?
Run the numbers first: identify the format, estimate the space, check the measured crack speeds. The math is the verdict — if it says months of GPU time even with your clues, trust it, and decide with facts instead of hope.
Sources & references
So what should you actually do?
-
1
Confirm the format
Identify the exact wallet file and version. That alone tells you — within minutes — whether your case is feasible: the format fixes the encryption and the speed.
-
2
Build the clue set
Everything you remember, in writing — the memory guide shows how to generate candidates.
-
3
Back up the file read-only
Never work on the original. Do this before anything else.
-
4
Run the tools yourself
Test your candidate list offline, free, on your own machine — follow the btcrecover walkthrough or the hashcat walkthrough.
Keep going — it’s all free
Not sure what you lost or what’s possible?
Take the quick diagnosis for a first verdict — it stops early when the answer is already clear — then work the search checklist. Everything here is informative, runs offline, and is free to use.