galaxy-cli — Galaxy bioinformatics command-line client

galaxy-cli is a Python CLI and REPL for automating the Galaxy bioinformatics platform from the shell. It wraps the Galaxy REST API for histories, datasets, collections, tools, workflows, jobs, and libraries.

This project was initially generated with cli-anything and then refined into the standalone galaxy-cli package.

pip install galaxy-cli

What can galaxy-cli do?

galaxy-cli is a Galaxy bioinformatics CLI for users who want a Galaxy command-line client that works naturally in shell scripts and reproducible Galaxy analyses. You can use it to:

• run Galaxy workflows from the command line and submit individual Galaxy tools
• upload and download datasets
• create and manage histories and collections
• inspect jobs, logs, tools, and workflow metadata
• automate Galaxy workflows for reproducible Galaxy bioinformatics analyses

Prerequisites

• Python 3.9+
• A running Galaxy server
• A Galaxy API key from your target instance

Installation

Install from PyPI with uv or pip:

uv tool install galaxy-cli

pip install galaxy-cli

For local development from this repository:

python3 -m pip install .

Verify installation:

which galaxy-cli
galaxy-cli --version

AI Agent Skill

The package includes a galaxy-cli skill for Codex and Claude Code so agents can use the CLI without rediscovering command syntax.

# Install for Codex: ~/.codex/skills/galaxy-cli/SKILL.md
galaxy-cli skill install --agent codex

# Install for Claude Code: ~/.claude/skills/galaxy-cli/SKILL.md
galaxy-cli skill install --agent claude

Use galaxy-cli skill path to find the packaged skill, or galaxy-cli skill install --target-dir /path/to/skills for another agent or project-level skills directory.

Configuration

Set your Galaxy server URL and API key:

export GALAXY_URL=https://usegalaxy.org
export GALAXY_API_KEY=your-api-key

galaxy-cli config test

Session state is stored in ~/.galaxy-cli/session.json. It is intended for a single active writer; for parallel automation or multiple concurrent agents, pass --history-id explicitly instead of relying on shared session state.

Usage

Documentation site: https://qchiujunhao.github.io/galaxy-cli/

galaxy-cli history list
galaxy-cli history create "My Analysis"
galaxy-cli tool search "bowtie" --limit 5
galaxy-cli workflow list
galaxy-cli workflow list | jq .
galaxy-cli --human config show

galaxy-cli defaults to compact JSON output. Use --human for human-readable terminal output.

Tool and workflow submissions validate obvious input mistakes before posting to Galaxy, including unknown input names, missing required data inputs, invalid dataset-vs-collection source prefixes, and simple select/boolean/numeric value errors. Use --dry-run-payload to validate and print the exact Galaxy POST body without submitting:

galaxy-cli tool run TOOL_ID --history-id HIST_ID -i input=DATASET_ID --dry-run-payload
galaxy-cli workflow run WF_ID --history-id HIST_ID -i 0=DATASET_ID --dry-run-payload

For large dataset uploads, use --upload-timeout to control the HTTP POST separately from the upload job wait. GALAXY_CLI_REQUEST_TIMEOUT controls regular API request reads, and GALAXY_CLI_UPLOAD_TIMEOUT controls uploads.

Run the REPL:

galaxy-cli

Tests

python3 -m pytest galaxy_cli/tests/ -v

Releases

Releases are published from GitHub Releases through PyPI Trusted Publishing. See RELEASE.md for the one-time PyPI setup and release checklist.