▚PANDA//VAULT
privacy is not a feature — it's the foundation

about

PANDA//VAULT is a free toolkit of zero-knowledge, privacy-first utilities — encrypted chat, anonymous file drop, burn-after-reading paste, disposable mail, a private forum, an offline encryptor, PGP signing and verification tools. No accounts. No tracking. No logs. The guiding rule is simple: you can't leak, lose, or be forced to hand over what you never hold.

our ethos

Privacy should be the default, not a paid upgrade or a buried setting — and it should be available to everyone, anonymously, with nothing to sign up for. We build tools that hold as little as physically possible about you, and we'd rather a feature do less than quietly know more than it claims.

Where a tool encrypts your data, it is encrypted in your browser before anything is sent — the server only ever sees ciphertext it cannot read. We don't log IP addresses. We don't set tracking cookies. We make no third-party requests, so no outside company ever sees that you were here. And almost everything self-destructs on a timer.

what sets us apart

No accounts, ever

Nothing to register, no email, no profile. There is no user database to breach because there are no users — only ephemeral, encrypted data.

We log no IP addresses

Even abuse rate-limiting is done with a one-way keyed hash that stores no address and rotates daily, so buckets can't be reversed or linked across days.

Zero third-party requests

No CDNs, no fonts from Google, no analytics, no ad pixels. Every asset is self-hosted, so no external service ever observes a visitor.

Identity & metadata encrypted too

Not just the message body — handles, room names and precise timestamps are bundled inside the ciphertext or hashed, so the server can't see who, where, or exactly when.

Readable, auditable code

The client code on this site ships unobfuscated. Open the source, read it, verify our claims for yourself — security through transparency, not obscurity.

Honest about limits

We only call something "end-to-end encrypted" or "zero-knowledge" when it literally is. Each tool states plainly what it does — and doesn't — protect.

technology & code hardening

Every tool is built on vetted, standard cryptography (the browser's Web Crypto API and the audited OpenPGP.js library) and shipped behind layered server hardening:

AES-256-GCM authenticated client-side encryption
PBKDF2-SHA256 100k–250k-iteration key stretching
RSA-4096 / Curve25519 in-browser PGP, self-hosted
Keys in the URL fragment never sent to the server
Ciphertext-only storage random IDs, non-executable dirs
HMAC integrity seals tamper-evident stored blobs
TLS 1.3 + HSTS preload encrypted, downgrade-proof transport
Strict CSP no inline scripts, no eval, own origin only
Subresource Integrity sha384-pinned, content-hash cache-busted
COOP / CORP + nosniff cross-origin isolation, no MIME sniffing
Coarsened timestamps hour-floored; no per-second metadata
Self-destruction auto-expiry, size & download caps
Build-time leak guard blocks any infra detail from shipping
Encrypted deploys FTPS/TLS 1.3, credentials sealed at rest

how we design

Threat-model first. We define what each tool defends against — and state honestly what it doesn't — before writing a line of code.
Least knowledge. We design so the server holds the minimum: ciphertext, opaque IDs, coarse timestamps. The best-protected data is data we never collect.
Defense in depth. No single control stands alone — strict CSP and output encoding, unguessable IDs and access checks, encryption and short retention.
Encrypt identity and metadata, not just the payload. Who, where and when are as revealing as the message, so they go inside the envelope too.
Verify, don't assert. Every cryptographic claim is proven with real encrypt → decrypt round-trips (including wrong-key failures) before it ships.
Accurate claims only. Overstating security on a security site is itself a security failure. We measure every feature against an internal design & security playbook.

⚠ Honest by default

Security marketing usually rounds up. We round down. If a tool isn't truly end-to-end encrypted, we won't pretend it is — for example, our disposable email is anonymous and ephemeral, but inbound mail arrives in plaintext over the internet, so the receiving server necessarily sees it; we label it receive-only and anti-tracking, never "zero-knowledge."

Likewise, no website can defeat a compromised device, and passphrase-based encryption is only as strong as the passphrase. We tell you these things up front, because trust is built on what a tool can't do as much as what it can.

on the roadmap

planned · not yet available

Everything here runs in the browser today. Next, we plan to bring the same zero-knowledge toolkit to native desktop applications for Windows and Linux — so the tools work offline-first, outside any browser, with the same no-accounts, no-tracking principles.

  • Native Windows and Linux builds of the encryptor, signing and verification tools
  • Fully offline operation — no network required for the local-only tools
  • Portable mode that runs without installation
  • The same readable, auditable code we ship on the web

These are intentions, not promises of a date — we'd rather ship them right than ship them fast.

PANDA//VAULT · about · acceptable use ← back to vault