transformerlab-cli 0.0.21

Transformer Lab CLI

Transformer Lab CLI

Install

uv tool install transformerlab-cli

Usage

 Usage: lab [OPTIONS] COMMAND [ARGS]...

 Transformer Lab CLI

╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --format        TEXT  Output format: pretty or json [default: pretty]        │
│ --help                Show this message and exit.                            │
╰──────────────────────────────────────────────────────────────────────────────╯
╭─ Commands ───────────────────────────────────────────────────────────────────╮
│ version   Display the CLI version.                                           │
│ config    View or set configuration values.                                  │
│ status    Check the status of the server.                                    │
│ login     Log in to Transformer Lab.                                         │
│ logout    Log out from Transformer Lab.                                      │
│ task      Task management commands                                           │
│ job       Job management commands                                            │
╰──────────────────────────────────────────────────────────────────────────────╯

Command Reference

For a full list of all commands with detailed options and example usage, see COMMANDS.md.

Claude Code Skill

Want Claude Code (or other AI coding agents) to manage Transformer Lab for you? Install the skill:

npx skills add transformerlab/transformerlab-app --skill transformerlab-cli

This teaches your AI agent how to use the lab CLI to check job status, stream logs, download artifacts, queue tasks, manage providers, and more. See .agents/skills/transformerlab-cli/ for the full skill definition.

Development

Run (DEV)

uv run src/transformerlab_cli/main.py

Build Locally

Option A — Global install (simple)

uv tool install . --force --reinstall

This installs lab into ~/.local/bin/ so it's available from any shell.

Option B — Editable install into an active venv (recommended for iterative dev)

# Activate the venv you want lab installed into, e.g.
source ~/.transformerlab/envs/general-uv/bin/activate

# From the cli/ directory:
uv pip install -e .

lab will live at <venv>/bin/lab and only resolve while that venv is active. Code changes take effect without reinstalling.

⚠️ Gotcha: if a cli/.venv directory exists, uv pip install targets it instead of your active $VIRTUAL_ENV, and which lab will come up empty. Either delete it (rm -rf cli/.venv) or force the target explicitly: uv pip install -e . --python "$VIRTUAL_ENV/bin/python".

Debug the Job Monitor

Install editable into your active venv (see Option B above), then:

uv run textual run --dev src/transformerlab_cli/commands/job_monitor/job_monitor.py

and then in another window:

textual console -x SYSTEM -x EVENT -x INFO

Run Textual in Browser

uv run textual serve src/transformerlab_cli/commands/job_monitor/job_monitor.py