LastPass, 1Password, Bitwarden and Proton Pass are established products built around provider-hosted encrypted vaults. KeePass takes a local-file approach, but does not use BlindLock's built-in steganographic carrier and device binding. BlindLock takes a different approach to both models. Here is what that changes.
| Criterion | BlindLock | LastPass | 1Password | Bitwarden | Proton Pass | KeePass |
|---|---|---|---|---|---|---|
| Vault storageWhere your encrypted vault physically lives | PNG carrier; separate file-vault containers | Cloud | Cloud | Cloud (self-host possible) | Cloud | Local only (KDBX file) |
| Account & loginWhat unlocks the account and vault | No vault cloud account — carrier + password + device | Email + master password | Email + password + Secret Key | Email + master password (passkey login available) | Proton account (email + password) | No account — password/key file |
| Steganographic vaultVault concealed in an ordinary-looking file (deniability depends on threat model) | ✓ — PNG carrier; optional decoy region | — | — | — | — | — |
| Long-term ciphertext riskHarvest-now / decrypt-later exposure if vault ciphertext is bulk-stolen | No central vault DB to harvest; at rest: 256-bit authenticated encryption (AES-256-GCM + XChaCha20-Poly1305) | Central encrypted vaults can be bulk-stolen offline | Central vaults; PQ hybrid TLS on web transport (not vault-at-rest) | Central encrypted vaults (self-host possible) | Central encrypted vaults | Local file only — no provider vault server |
| Hardware-bound keysKey material anchored in platform security hardware | TPM 2.0 / Secure Enclave / Android StrongBox or TEE where available | — | Device-linked (cloud-backed) | — | Device-linked | No hardware sealing (KeePassXC: quick unlock only) |
| Hardware-bound unlockVault key material is sealed to platform security hardware on the authorised device | ✓ — sealed to TPM 2.0 / Secure Enclave / mobile hardware root; PIN gates unlock in the app | — | — | — | — | — |
| Central customer-vault databaseWhether the provider stores encrypted customer vaults | None — vault contents are never sent to BlindLock | Server-side vault | Server-side vault | Server-side (self-host possible) | Server-side vault | Local file; no provider vault server |
| Key derivationWhat a single brute-force guess costs the attacker (time × memory) | Argon2id, memory-hard (512 MiB–4 GB by device class) | PBKDF2 — not memory-hard, GPU-friendly | PBKDF2 + SRP — not memory-hard | PBKDF2 (600k) default for many accounts; Argon2id available (~32–64 MiB) | bcrypt + SRP — low memory cost per guess | KeePassXC: Argon2 (~64 MB default) / KeePass 2.x: AES-KDF by default |
| Open cryptographic layerWhether crypto primitives or the full app are publicly inspectable | libcrux-backed AEAD primitives (verified cores where available); app not open source | Closed | Closed | Open source | Partially open | Open source |
| TOTP 2FA built inNo second authenticator app needed | Built in — codes stay inside the vault | Free plan includes TOTP; advanced/hardware MFA on paid tiers | ✓ | Paid tier (Premium) | Paid tier (Plus) | KeePass 2.x: third-party plugin required / KeePassXC: built in |
| Pricing modelOne-time vs. recurring revenue | Lifetime desktop licence (pay once; one active desktop; moves to a new PC; iOS & Android always free, for everyone) | Subscription | Subscription | Subscription (free tier) | Subscription (free tier) | Free (open source) |
| Public breach historyMajor disclosed vault exposure | No central vault service to breach | 2022 — encrypted vault backups stolen (~30M reported); cracking linked through 2025 | No vault exfiltration | No vault exfiltration | No vault exfiltration | No central provider vault |
If someone steals an encrypted vault file, they can try passwords offline on graphics cards. Low-memory methods let them try many passwords at once. High-memory methods force them to go much slower. A strong password still matters either way.
| Product | Key derivation (typical default) | Memory per guess | Offline guessing on a GPU |
|---|---|---|---|
| LastPass / 1Password | PBKDF2 | Very low | Many parallel guesses |
| Bitwarden | PBKDF2 or Argon2id (account-dependent) | Very low to medium | Many parallel guesses on PBKDF2; fewer if Argon2id is enabled |
| Proton Pass | bcrypt | Very low | Many parallel guesses |
| KeePass 2.x | AES-KDF (default) | Very low | Many parallel guesses |
| KeePassXC | Argon2 (~64 MB default) | Medium | Hundreds of parallel guesses on a 24 GB GPU |
| BlindLock | Argon2id (512 MiB–4 GB by device class) | High | Far fewer parallel guesses; normal unlock also requires the authorised device (recovery is a separate path) |
You pay the memory-hard cost once when unlocking. An attacker pays it for every guess. Values above are typical shipping defaults; some products let users raise them. BlindLock’s vault password KDF scales with the device class that created the vault; separate file-vault envelopes stay at the 512 MiB floor.
The matrix above is specific. These are the patterns behind it.
A zero-knowledge cloud manager still transmits an encrypted vault to provider infrastructure. BlindLock never transmits vault contents to our servers and operates no central customer-vault database or vault cloud account. Licence and version checks remain strictly separate from your secrets.
Encrypted vaults stolen today can be stored and attacked later. BlindLock reduces that exposure by design: there is no central vault database to harvest, and resting vault files use 256-bit authenticated encryption, generally regarded as beyond practical quantum attack for confidentiality under current estimates. That reduces risk; it is not a promise that every future attack is impossible.
An attacker who obtains a cloud vault can often attack it offline on any machine. With BlindLock, normal unlock needs the carrier file, the password, and the authorised device. Optionally, a security key (YubiKey, Google Titan, or SoloKey) adds a fourth factor. Recovery is a separate, deliberate path.
Lifetime pricing opens soon — phases from €59.