This article may contain affiliate links. If you buy through them we may earn a commission at no extra cost to you.
A password manager encrypts your credentials on your device using AES-256, stores the encrypted vault locally or in the cloud, and decrypts it only after you supply the correct master password. The company hosting your vault cannot read it—not because of policy, but because the decryption key never leaves your device.
What Gets Stored and How
When you save a login, the manager encrypts the username, password, and URL together as a single vault entry. Most managers also store:
- Secure notes (arbitrary encrypted text)
- Payment card numbers and CVVs
- Identity fields (name, address, phone) for form auto-fill
- Passkeys and TOTP codes, on managers that support them
The password generator typically defaults to 16–20 random characters mixing letters, numbers, and symbols. That length puts a brute-force attack well beyond practical reach with current hardware—a 20-character random password has more combinations than there are atoms in the observable universe.
How Encryption Protects Your Data
Encryption happens in two distinct stages, and conflating them causes most of the confusion about how secure these tools actually are.
Encryption at Rest
Your vault is encrypted with AES-256, a symmetric cipher. To generate the decryption key from your master password—which is just a human-readable string—the manager runs it through a key derivation function (KDF). Bitwarden uses PBKDF2 with a default of 600,000 iterations; 1Password uses Argon2id. The iteration count matters: each additional round makes a brute-force attack proportionally slower without affecting your unlock time noticeably.
Your master password is never stored anywhere—not on your device, not on the company’s servers. What’s stored is a derived key hash. If an attacker steals the vault file, they must guess your master password and run it through the same KDF to check whether it’s correct. At 600,000 PBKDF2 iterations, even a dedicated GPU cluster tests far fewer guesses per second than against a simple hash.
Encryption in Transit
When your vault syncs to the company’s servers, it travels over TLS. Critically, the vault is already encrypted before it leaves your device, so the TLS layer is a second envelope around an already-sealed one. A network intercept yields ciphertext that’s no more useful than the copy on the server.
End-to-end encryption means your password manager company cannot read your passwords, even if their servers are breached.
Local vs. Cloud Storage
The storage model changes your threat surface, not the encryption strength.
Cloud-Synced
Your encrypted vault lives on the provider’s servers and syncs automatically to every device. The practical advantage is resilience: a lost or broken phone doesn’t cost you your passwords. The trade-off is that the provider’s servers become a high-value target. The LastPass breach of 2022 demonstrated this—attackers exfiltrated encrypted vaults. Users with strong, unique master passwords remained protected; those with weak ones faced real risk as attackers could run offline brute-force attempts at their own pace.
Local-Only
KeePass and its forks store the vault as a single encrypted file on your device. There’s no company server to breach. The cost is manual sync—you move the file between devices yourself via USB, a personal cloud folder, or a local network share. If your only copy is on a drive that fails, the vault is gone.
For most people, the cloud-synced model with a zero-knowledge provider is the better trade-off. Zero-knowledge means the provider’s architecture makes it technically impossible for them to decrypt your vault, regardless of a court order or a breach.
Authentication: Master Password and Beyond
Your master password is the single point of failure, so its strength has an outsized effect on your actual security. A 12-character random passphrase is a reasonable floor; 16+ characters is better. Dictionary words, names, and dates are poor choices because offline brute-force attacks use those patterns first.
Two-factor authentication (2FA) adds a second requirement to unlock your account on a new device. Options differ meaningfully in what they protect against:
- TOTP apps (Authy, Google Authenticator): Protect against remote account takeover. Don’t help if malware is already on your device.
- Hardware security keys (YubiKey, Google Titan): Resist phishing because the key signs a challenge tied to the exact URL—a fake site gets a different challenge and can’t authenticate. The strongest option for account-level protection.
- Biometric unlock: Unlocks the cached vault on your device using your fingerprint or face. The biometric data never leaves your device and never reaches the provider. Convenient but only relevant to local device access.
How Auto-Fill Works Without Exposing Passwords
The browser extension monitors for login forms by scanning for password-type input fields and checking the page’s URL against your vault entries. URL matching is intentional: if you’re on paypa1.com instead of paypal.com, no match is found and auto-fill doesn’t trigger—a passive defense against phishing that works without any action from you.
When a match is found and you approve the fill:
- The extension requests the decrypted credential from the local vault process.
- The vault decrypts the entry in memory.
- The plaintext password is injected directly into the form field.
- The decrypted string is discarded from memory; it’s never written to disk, the clipboard, or browser history.
This entire sequence runs on your device. The provider’s servers are not contacted during a fill.
Vulnerability Points and Mitigations
Three attack surfaces account for most real-world compromises.
Malware on Your Device
A keylogger can capture your master password as you type it. Memory-scraping malware can, in theory, read decrypted passwords before they’re cleared. Hardware 2FA limits the damage—an attacker with your master password still can’t log into your account on a new device without the physical key. Keeping your OS patched and avoiding untrusted software reduces exposure.
Company Server Breaches
Attackers gain encrypted vault files. With a strong master password and a provider using high-iteration KDFs, those files are computationally impractical to crack. The risk concentrates on users who reused weak master passwords. Choosing a provider with published third-party security audits (1Password, Bitwarden, and Dashlane all publish them) gives you evidence of their implementation, not just their claims.
Browser Extension Exploits
Extensions run with elevated browser permissions. A malicious update or a cross-site scripting vulnerability in the extension could expose credentials. Keeping your browser and extensions updated closes known CVEs quickly. Reviewing extension permissions—a password manager extension shouldn’t need access to your microphone or camera—catches obvious overreach.
Comparing Password Manager Approaches
| Feature | Cloud-Synced (Bitwarden, 1Password) | Browser-Integrated (Chrome, Safari) | Local-Only (KeePass) |
|---|---|---|---|
| Multi-device sync | Automatic | Limited to same browser ecosystem | Manual |
| Zero-knowledge encryption | Yes (reputable providers) | No—Google and Apple can access data | Yes (local only) |
| Company breach risk | Low (encrypted vault only) | Higher (data stored less securely) | None (no servers involved) |
| Setup complexity | Easy | Very easy | Moderate |
| Mobile access | Full app support | Limited to same-vendor devices | Requires manual sync setup |
Our verdict
Cloud-synced managers with zero-knowledge architecture—Bitwarden and 1Password are the clearest examples—give you automatic sync, strong encryption, and published audits. Browser-integrated managers trade encryption rigor for convenience; Google and Apple have the technical ability to access what you store there. Local-only managers eliminate third-party risk entirely but require you to manage your own backup and sync. For specific product recommendations, see our best password manager for personal use review, and for a detailed look at the security gap between dedicated managers and browser storage, see password manager vs. browser storage.
Frequently asked questions
Is it safe to store all passwords in one place?
The centralized risk is real but manageable. A strong, unique master password combined with hardware 2FA makes the vault significantly harder to compromise than the alternative—reused passwords stored in a browser. The practical comparison isn’t “one place vs. many places”; it’s “one well-encrypted place vs. weak passwords reused across dozens of sites.”
Can password managers be hacked?
Provider servers can be breached—LastPass in 2022 is the most cited example. Attackers obtained encrypted vaults but not the master passwords needed to decrypt them, because those were never stored. The residual risk falls on users whose master passwords were weak enough to crack offline. Breaches can also expose your account email address and any unencrypted metadata (like URL lists), which can be used in targeted phishing. A provider with independent security audits and a track record of transparent breach disclosure is worth the premium over one without.
What happens if you forget your master password?
With most zero-knowledge providers, the vault is permanently inaccessible—there’s no server-side copy of your key to recover. Some managers offer an emergency recovery kit (a printed code you generate at setup) or a trusted-contact recovery option, but both introduce a secondary attack surface. Store your master password in a physically secure location, separate from your devices.
Do password managers work offline?
Cloud-synced managers cache an encrypted copy of your vault on each device, so unlocking and reading passwords works without an internet connection. New entries or edits queue locally and sync when you reconnect. Local-only managers like KeePass work entirely offline by design. Browser-integrated managers vary—Chrome’s password manager requires connectivity to sync but can read cached credentials offline.
Can websites detect that you’re using a password manager?
Websites can’t identify the software filling a form, but they can detect the autocomplete attribute behavior and set their own fields to resist it. Some banking and government sites explicitly block auto-fill, requiring you to type credentials manually. This is a site-side decision, not a limitation of the manager. The manager itself leaves no detectable fingerprint that distinguishes it from a user typing quickly.
Dev has benchmarked laptops and phones for a decade and still reads the changelog before updating.
