Onehouse CLI

The Onehouse CLI (oh) is a command-line tool for managing all Onehouse resources from your terminal — clusters, lakes, databases, tables, flows, jobs, transformations, and more.

It is a thin layer over the Python SDK Resources client: every command generates a SQL statement and posts it to the Onehouse control-plane API, then streams the result to your terminal as a formatted table or JSON.

pip install onehouse-cli

Requires Python 3.9+. The binary is oh (alias: onehouse-cli).

What's in this section

Page	What it covers
Installation & configuration	Installing, running oh-configure, credential files, profiles, environment variables, AI agent setup
Global options	--profile, --timeout, --output, --show-sql, async mode
Command reference	Every command group and subcommand with all flags

Quick example

# Configure credentials once
oh-configure

# List all clusters
oh cluster show

# Create a lake
oh lake create my-lake \
  --type MANAGED \
  --bucket-path s3://my-bucket/lake/ \
  --default-services-cluster my-cluster

# Create a flow
oh flow create web-events \
  --source kafka-prod \
  --lake my-lake --database events --table-name clicks \
  --write-mode MUTABLE \
  --cluster my-cluster

# Preview the SQL sent to the API without executing
oh cluster show --show-sql

AI agent integration

The CLI ships with a built-in skill for AI coding assistants. Install it once and your agent learns every oh command, flag, and workflow — including how to write and run Spark jobs:

oh install agent-skill --claude    # Claude Code / claude.ai/code
oh install agent-skill --cursor    # Cursor IDE
oh install agent-skill --codex     # OpenAI Codex CLI
oh install agent-skill --all       # all of the above

After installation, ask your agent things like:

• "Show me all my clusters" → runs oh cluster show
• "Write a PySpark job that aggregates this data and run it on Onehouse" → writes the script, uploads it, and creates the job

See Installation & configuration for details.