Skip to main content
Breeze CLI brings the Breeze Developer API to the terminal. Hear a voice seconds after installing, save takes as files, turn any command into a reproducible API request, and give scripts and AI agents a JSON-native way to create audio and narrated video.

​
Install

The installer prints a PATH hint when the install directory is not already available in your current shell.

​
Hear it first

breeze login opens your browser and stores a device-specific key locally — no key copy-paste. In an interactive terminal, breeze tts picks a voice for you when you do not pass one, and streams playback with a live waveform as soon as audio arrives.

​
Move into code

Every CLI call maps to the public /v1 Developer API — the same API behind the Breeze SDKs. When a take sounds right, ask the CLI for the equivalent request:
breeze curl prints a reproducible cURL command with a ${BREEZE_API_KEY} placeholder, ready to paste into a script or translate into an SDK call. See From CLI to API.

​
Built for agents

Add --agent to any command for machine-readable behavior: JSON results on stdout, JSON error envelopes on stderr, stable exit codes, no prompts, no spinners, no audio autoplay.
If Codex, Claude Code, or another coding agent is installing Breeze CLI for you, install the matching Breeze agent skills in the same step:
Use --with-skills claude for Claude Code. See Scripting and agents.

​
Choose a path

CLI quickstart

Install Breeze CLI, log in, and generate the first audio locally.

Text to speech

Generate speech, stream playback, save files, and run realtime conversation turns.

Voices

Browse voices, design new ones from a description, and clone from an audio sample.

Scripting and agents

Use JSON output, stable exit codes, capabilities discovery, and Breeze agent skills.

Command reference

The full command map, global flags, and environment variables.

Troubleshooting

Fix install, login, update, network, timeout, and Video Lab dependency issues.