Your Secrets Deserve to Be Invisible

In December 2022, LastPass disclosed that encrypted customer vault backups had been stolen — later reporting put the scale on the order of tens of millions of accounts (figures vary by source). Not a partial leak. Copied ciphertext, sitting with attackers. The only thing standing between those attackers and each vault’s contents was the strength of that user’s master password and key-derivation settings.

LastPass assured everyone that vaults were encrypted with AES-256. Technically true. But they also admitted that some vault data — including website URLs — was stored unencrypted. And for users with weak master passwords or older accounts using fewer encryption iterations? Their secrets were essentially gift-wrapped.

Two years later, researchers linked over $35 million in cryptocurrency theft directly to those stolen vaults.

This isn't a story about LastPass being uniquely careless. This is a story about a fundamental problem with how we think about password security. And it's a problem that's about to get much, much worse.

The cloud paradox

Here's the deal most password managers offer you: Give us all your secrets. We'll keep them safe. Trust us.

Every major password manager — 1Password, Dashlane, LastPass, even Proton Pass — operates on the same basic model. Your encrypted vault lives on their servers. You access it through their apps. They promise they can't read your data. And in most cases, they're telling the truth.

But "we can't read your data" and "your data is safe" are two very different statements.

When your vault sits on someone else's server, you inherit every risk that server carries: employee access, infrastructure vulnerabilities, government subpoenas, supply chain attacks, and the simple mathematical reality that a server holding tens of millions of vaults is a far more attractive target than your laptop.

And this is not theoretical. In June 2026, LastPass customer data leaked again — not from the vault, but through exactly this supply chain: an extortion group compromised the market-intelligence tool Klue, which was connected to LastPass's CRM, harvested the OAuth tokens it held, and made off with customers' names, email addresses, phone numbers, physical addresses, and support cases. The vaults stayed encrypted this time — customers' personal data did not. Trust your data to someone else's server ecosystem and you inherit every gap in it — right down to your providers' providers.

The value proposition of a cloud password manager rests on the encrypted vault remaining impractical to open. The 2022 LastPass theft put that assumption under sustained pressure: attackers could work on copied vaults offline, without returning to the provider. Password strength and key-derivation settings therefore mattered enormously.

Vaults protected by strong passwords and sound cryptography may remain secure, but the attackers still hold the ciphertext and can continue testing it offline. The provider cannot take that copied data back. That permanent exposure is the architectural risk.

They don't even need to be specialists anymore. In May 2026, a Bitcoin owner used an AI assistant to open a long-locked wallet holding roughly $400,000: the AI ran password-recovery tooling, fixed its bugs, and combed old backups — work that used to require an expert. He was recovering his own property, entirely legitimately. The point for password managers is simpler: once someone holds offline-attackable ciphertext, tooling keeps getting cheaper.

Long-term ciphertext risk

You've probably heard that large quantum computers could undermine important parts of today's cryptography. It is a future capability that influences security decisions made today — but the claims need to stay precise.

The strategy is called "harvest now, decrypt later": collect encrypted data now in the hope that future advances make parts of its protection easier to defeat. A cryptographically relevant quantum computer would directly threaten widely used public-key systems such as RSA and elliptic-curve cryptography. Symmetric authenticated encryption is affected differently and is not simply rendered useless overnight.

BlindLock’s first answer is architectural: there is no central customer-vault database to harvest at scale, and the resting password vault uses classical 256-bit dual AEAD — not a post-quantum wrap. Under known quantum attacks on symmetric crypto, that remains in a work-factor class generally treated as impractical. That is not a claim that quantum computers cannot affect cryptography at all.

NIST has finalised post-quantum standards so organisations can migrate carefully. BlindLock includes NIST ML-KEM-1024 and ML-DSA-87 in the stack for scoped uses (for example update and licence signatures and documented network paths). They do not make BlindLock a blanket “post-quantum vault.”

What if your vault was invisible?

Let's step back from cryptography for a second and think about physical security.

Imagine you own a safe. A really good safe — thick steel walls, a combination lock that would take centuries to crack. You put it in your living room. Everyone who visits your house can see it. They know you have something worth protecting. And if someone really wants what's inside, they have all the time in the world to work on that lock.

Now imagine something different. Instead of a safe, you take your valuables and hide them inside a painting on your wall. The painting looks completely normal. No one visiting your house knows there's anything hidden in it. There's nothing to attack because there's nothing to see.

That's steganography. And it's the core idea behind BlindLock.

Steganography hides data inside other files. With BlindLock, the complete encrypted password vault lives inside a PNG carrier on your device. The carrier still looks like an ordinary image file. The point is not magic: it is to avoid advertising an obvious vault file.

BlindLock operates no central customer-vault database and never transmits vault contents to our servers. Password, note and 2FA entries live in an ordinary-looking PNG carrier on your device. Opening it requires the correct carrier, your master password and the authorised device. If one is missing, the vault stays closed. You can also require a compatible security key, such as a YubiKey, Google Titan or SoloKey.

Concealment is not a cryptographic proof of plausible deniability. Carriers are designed without an obvious vault signature; that adds obscurity, not forensic immunity. Encryption remains the confidentiality boundary. VeraCrypt established hidden volumes for disks; BlindLock applies a related idea to a personal vault file — your secrets are encrypted and stored inconspicuously, not magically unfindable.

Trust math, not promises

Here's a question that should make you uncomfortable: How do you know your password manager's encryption actually works?

The honest answer for most software: You don't. You trust the company. You trust their engineers. You trust that their code review process catches bugs. You trust that they use well-known cryptographic libraries correctly.

But "we use AES-256" is not a security guarantee. It's a marketing statement. The encryption algorithm is the easy part. What matters is the implementation — and implementations have bugs. Heartbleed, the OpenSSL vulnerability that exposed millions of servers in 2014, wasn't a problem with the cryptographic algorithm. It was a problem with a few lines of C code that handled memory buffers.

This is why BlindLock builds on the libcrux / HACL* lineage for selected cryptographic cores.

HACL* (High Assurance Cryptographic Library) is designed with formal verification in mind. In plain terms: for the verified primitives, key properties such as memory safety and constant-time behaviour are mathematically checked — not merely “tested a lot.” That rules out entire bug classes inside those library cores. It does not automatically prove every line of BlindLock around them.

Those verified components strengthen the primitive layer. BlindLock’s own vault, steganography, UI and licence code still need testing, review and — over time — independent audit. We keep that boundary explicit, because your secrets deserve better than blurred marketing.

More than passwords

A password manager that only manages passwords is solving yesterday's problem.

Your digital life in 2026 includes two-factor authentication codes, recovery keys, secure notes, sensitive documents, and credentials for hundreds of services. Most of this data is scattered across apps, sticky notes, text files, and (let's be honest) your memory.

BlindLock is a unified local vault for the pieces that matter:

Passwords — Encrypted inside your PNG carrier, searchable, with local strength analysis. The core feature, done carefully.

2FA Codes — TOTP generation runs locally after unlock, without a separate authenticator app or cloud vault. The 2FA secrets stay encrypted inside your carrier and are never sent to BlindLock servers. A new unlock still requires the brief licence and version check used throughout the app.

Secure notes — Plain text and Markdown in the same encrypted carrier as your logins.

File vaults — Larger documents, media and archives live in separate disguised, encrypted container files; they are not stored inside the PNG password-vault carrier. BlindLock handles them through its integrated file explorer rather than mounting a virtual drive. The containers use Argon2id-hardened, 256-bit authenticated encryption and are designed without an obvious BlindLock header or vault signature. This reduces recognisability, but does not replace encryption or endpoint security.

Security keys — Optional fourth factor on desktop: compatible FIDO2 keys such as YubiKey, Google Titan and SoloKey.

Why lifetime, not subscription

Let's talk about money.

1Password costs roughly €44 per year. Dashlane charges about €55. Over five years, that's €170 to €275. Over ten years? €340 to €550 for a password manager.

BlindLock Lifetime costs a one-time €59 and up, with no expiry date or recurring subscription. All released updates and major versions are included. Licence entitlement is verified online when you unlock; security-critical releases may require a minimum version. Exactly 1,000 Lifetime licences are sold in three waves: the first 150 at €59, the next 350 at €79 and the final 500 at €99. After that, BlindLock is subscription-only.

We're doing this because BlindLock was built by a very small, independent team — and we intend to stay independent. No investors and no banks get to tell BlindLock what happens next, or how. Selling the 1,000 Lifetime licences puts the project on solid financial ground so the team can keep developing BlindLock. Buyers pay once for a licence with no expiry date or recurring subscription.

BlindLock combines steganographic carriers, hardware binding, scoped post-quantum stack components, and carefully chosen cryptographic primitives in one local-first vault — built for people who want more than conventional cloud password storage, without overselling what the product can guarantee.

The bottom line

The password manager market is full of products that are good enough. They encrypt your passwords, sync them across devices, and auto-fill forms. For most people, most of the time, they work fine.

But "good enough" has an expiration date. Cloud breaches keep happening. Offline copies of vault ciphertext do not expire. Long-term cryptographic risk is real — and the honest response is architecture plus precise scope, not slogans.

BlindLock was built for people who think ahead: local-first, concealed on disk, bound to their device — and clear about what concealment, encryption and formal methods can and cannot promise.

Your secrets deserve to be invisible.