1. BWEZE
  2. Security

How BWEZE keeps your work apart.

What protects your apps, data and wallet today, what is not done yet, and how to tell us about a problem.

Backups
Not running yet
Machine-readable
security.txt

Every workspace is walled off

Each workspace's apps run in their own namespace behind a default-deny network policy: nothing reaches them, and they reach nothing, unless a rule allows it. Private services answer only your other services, never the internet.

Secrets are encrypted

Environment variables are encrypted at rest. Workspace secrets are envelope-encrypted under a key-encryption key, and the control plane refuses to start without a valid one, so it cannot run with secrets in the clear.

Apps and the dashboard never share a domain

Your apps are served on bweze.app; the dashboard and the API live on bweze.com. The two never share a parent cookie domain, so code running in an app cannot read a dashboard session. Every app is served over HTTPS.

The cluster pulls, it is never pushed

The control plane holds no credentials for the cluster that runs your apps. An agent inside the cluster pulls the state it should be in and applies it, so a compromised dashboard cannot reach into the cluster directly.

Money is kept by the database

The billing ledger is append-only, every transaction balances, and idempotency keys stop a charge from being taken twice. Postgres enforces all three with constraints, not application code. The wallet never runs on credit.

API tokens you can narrow

An API token can be limited to certain actions, to certain projects, backends or zones, to an IP range, and to a window of time. A request outside those limits is refused with its own error code, so you can tell why.

What is not done yet

  • Backups are not running yet. Until the nightly backup job runs, keep your own pg_dump of anything you cannot lose.
  • The beta runs in one data center, so it has one place that can fail.
  • There is no SLA. We will not offer one until we can measure it.

Report a vulnerability

Email [email protected] with what you found and the steps to reproduce it. Please give us time to fix it before you share it.

  • Test only against your own BWEZE projects. Never attack another customer's project.
  • Do not run automated scanners without telling us first, and then only against your own project.
  • Out of scope: denial of service, social engineering, attacks that need physical access or a man in the middle, and missing headers or flags with no clear attack.

Questions that are not vulnerabilities go to [email protected].

Found something? Tell us first.

[email protected]