Skip to content
Context Code Context Code

Security and privacy

Context Code is built on one premise: your data is yours and it stays on your machine.

Where your data lives

Section titled “Where your data lives”

All your data — chats, history, settings — is stored on your disk, never on a Context Code server. The only thing that leaves your machine is the requests to the AI providers you configure (and if you use local models only, not even that).

Encryption

Section titled “Encryption”
  • The database is encrypted end to end with SQLCipher.
  • Conversation text carries its own AES-256-GCM encryption on top of that.
  • Encryption keys are kept in the operating system keychain — Credential Manager on Windows, Keychain on macOS, Secret Service on Linux — not in loose files.

Control over the agent

Section titled “Control over the agent”
  • The approval modes define what the agent can do without asking you, from “chat only” to “full access”.
  • The optional sandbox mode confines the agent to the project folder and blocks access to sensitive system commands.
  • The app never elevates permissions without asking. A command that needs administrator reaches you as a card with the full command — in every approval mode — and then goes through the system’s UAC dialog. You can switch it off entirely in Settings → Security; see Running as administrator.

Everything remote is closed by default

Section titled “Everything remote is closed by default”
  • The messaging channels and web access ship disabled; they only turn on with an explicit action from you.
  • Every channel requires an allowed-users list: without it, the bot answers nobody.
  • Web access is protected by a token with a configurable expiry and is only visible on your local network, unless you explicitly create a public URL.

No telemetry

Section titled “No telemetry”

Context Code sends no telemetry to its servers. There is no hidden analytics and no “anonymous usage statistics”.