Is it safe for CI and AI agents?

Yes — non-interactive use is a first-class citizen. A Terminalwire CLI is just a command line, so CI jobs, scripts, and AI agents like Claude Code and Codex run it the same way a human does: type the command, read the output, check the exit code.

Two design decisions make that dependable:

• The license check never gets in the way. It never blocks a command, never prompts, never gates output, and never changes an exit code. If our licensing service were unreachable entirely, your CLI keeps working — it fails open, always.
• Deterministic behavior across versions. The client auto-updates from us and is tested against every server implementation with one conformance suite, so the command that worked in CI yesterday works after the update tonight.

Your automation authenticates the same way your humans do — through your web app’s existing auth — so there’s no separate token store to manage or leak.