Advanced

Advanced Pipekit CLI usage. Docker container patterns, non-interactive login, token storage, log-level flags.

For the standard install and command flow, see Install and Commands. This page covers patterns you'll reach for in CI, containers, or multi-account setups.

Non-interactive login

Log in by passing credentials as arguments:

pipekit login -u username -p password
# or
pipekit login --user username --password password

Or via environment variables:

export PIPEKIT_USERNAME=username
export PIPEKIT_PASSWORD=password
pipekit login

This is the path the Docker container section and CI runners take.

Token storage location

By default, the CLI writes the access token to ~/.pipekit/token. Set PIPEKIT_CONFIG_DIR to point at a different directory; the CLI then reads and writes $PIPEKIT_CONFIG_DIR/token:

export PIPEKIT_CONFIG_DIR=/tmp/pipekit-ci
pipekit login -u username -p password

Use this when:

• Running the CLI in CI, containers, or integration tests where each invocation needs an isolated token store.

• Running multiple agents (e.g. MCP servers) against different Organizations from the same host.

Logging flags

The CLI accepts --log-level and --log-format on every command:

Flag	Values	Default
--log-level	debug, info, warn, error, fatal	info
--log-format	text, json	text

JSON logging is the right choice when feeding CLI output into a log aggregator.

Docker container

The basics of running pipekit13/cli are in Install > Docker container. This section covers more involved patterns.

Used within a workflow

Embed the CLI inside an Argo Workflow step:

Used within another container

Copy the CLI binary into a container you control. This example adds it to a Jupyter notebook image:

MCP server pointer

The CLI also hosts an MCP server. For LLM client integration, see AI > Setup.