How CloudBox is built

A quick mental model so the rest of the docs make sense.

Storage that never goes down

Writes are mirrored to more than one storage node (a primary Synology array plus a backup). A write succeeds while any node is up; if one is unreachable the write is queued and replayed when it returns. You never see an outage, and reads fail over automatically.

Zero-knowledge by default

Free transfers and your private drive are encrypted in your browser before anything leaves your device. The server only ever stores ciphertext, and the decryption key rides in the share link's #fragment — it never reaches our servers. That's why a download link can't be re-sent from the server: only the browser that created it holds the key.

A clearly-labelled AI zone

An AI can only work with what the server can read, which is the opposite of zero-knowledge. So AI access is limited to a separate, opt-in AI folder — server-readable, clearly labelled, and the only thing the connector ever exposes. Your private files and transfers stay zero-knowledge. See AI connectors.

Accounts & keys

• Account — your login: dashboard, drive, plan, usage. Free or paid.
• API key — a key ID (cbk_…) + secret (cbs_…) your app uses. Live or test.
• Sub-account — a client or teammate you create under your master account. See Teams.