CLI

layer is Layer’s command-line interface. It reaches the same models, workspaces, and Creative Unit balance as the app, from a terminal, a Makefile, or a CI job.

uv

uv tool install layer

pipx

pipx install layer

pip

pip install layer

Then sign in and generate:

layer login
layer generate image -p "a cyberpunk market stall, isometric"

The first run picks the recommended model for the modality, prints what it will cost, waits for the result, and writes the file into your working directory.

InstallInstallers, supported Python versions, and upgrading.

AuthenticationBrowser sign-in, device codes, and API keys for CI.

ConfigurationProfiles, layer.toml, and the environment variables that override them.

Generating assetsImage, video, mesh, and audio — every flag, and how runs are priced.

ModelsList what your workspace can run, and inspect what a model accepts.

ScriptingJSON output, the stdout/stderr split, and the published exit codes.

Command surface

Command	What it does
layer login / logout / whoami	Sign in, sign out, and show who you are signed in as
layer use <workspace>	Choose the workspace subsequent commands act on
layer generate image|video|mesh|audio	Run a generation
layer generate status <id>	Read back a run submitted with --no-wait
layer base-models list / show	The catalogue your workspace can run, and one model’s inputs

Global options apply to every command: --profile, --workspace/-w, --project, --api-url, --json, --quiet/-q, and --version.

Other ways to drive Layer

The CLI is for you and your scripts. If the caller is an AI agent rather than a shell, the MCP server gives it tools instead, and the Agent Skills teach it the procedures. For your own services, there is the REST API.