Running Pipelines
Basic usage
Run a pipeline by name:
pipe deployPipe resolves the name through the alias chain, then Hub, then local files at ~/.pipe/files/.
Variable overrides
Pass variables as positional arguments after the pipeline name:
pipe deploy env=production tag=v1.2.0These override any values in the pipeline’s vars section. See Variables & Templating for the full precedence chain.
Verbosity levels
| Flag | Level | Behavior |
|---|---|---|
| (none) | Warn | Compact mode — minimal output, suppresses info noise |
-v | Info | Verbose — shows step start/finish, variable resolution |
-vv | Debug | Debug — includes internal details, env construction |
pipe deploy -v # verbosepipe deploy -vv # debugCompact UI
At the default verbosity (no -v flag), Pipe shows a compact status display: a spinner with the current step name, replaced by a success or failure indicator when the step completes. This keeps the terminal clean for pipelines with many steps.
Logs
Run logs are written to ~/.pipe/logs/. Each run produces a log file identified by the run ID.
Log rotation
Control how many log files to keep with the PIPE_LOG_ROTATE environment variable:
export PIPE_LOG_ROTATE=10 # keep last 10 log filesSet to 0 to disable rotation (keep all logs). Unset or invalid values use the default.
State files
Each run saves its state to ~/.pipe/state/<pipeline>/. State files record which steps completed, their exit codes, and outputs (unless marked sensitive). This is what enables --resume — see Resuming Failed Runs.
State rotation
Control how many state files to keep with the PIPE_STATE_ROTATE environment variable:
export PIPE_STATE_ROTATE=5 # keep last 5 state files