Working with this CLI via an AI agent? Agents can run
datumctl directly. Use -o json and the structured errors described in Output and scripting, and install Datum’s Skills so the agent knows the canonical commands for each resource. For an MCP client instead of a shell, see Datum MCP.datumctl is the command-line tool for working with Datum Cloud. It exposes a consistent resource model — the same get, apply, describe, and delete verbs across every resource type — so both developers and AI agents get a fast, predictable way to interact with Datum.
Ready to play? Install
datumctl and get started.Find your way around
Authentication
Sign in, use service accounts for automation, and manage sessions and tokens.
Working with resources
Read, change, and safely preview changes to Datum Cloud resources.
Contexts & scoping
Control which organization and project your commands act on.
Compute
Deploy and manage containerized workloads with the compute plugin.
Activity & audit
Investigate who changed what, review events, and follow the activity feed.
Interactive console
Browse and monitor resources in a full-screen terminal UI.
AI & MCP
Ask about your resources in natural language, or connect AI clients over MCP.
CI/CD
Install and authenticate
datumctl in GitHub Actions and other pipelines.Plugins
Extend the CLI with additional commands from plugin catalogs.
Product plugins
Each Datum product adds its own commands todatumctl as an official plugin. Install the ones you use, and run datumctl plugin search to see them all:
datumctl --help lists the plugins you have installed, and the plugins in the catalog that you haven’t installed yet.
How it works
At a high level,datumctl does three things:
- Authenticates you - Opens a browser login, stores credentials securely, and refreshes tokens automatically.
- Talks to the Datum API - Sends requests to create, read, update, and delete resources with authentication handled behind the scenes.
- Formats output - Displays results as tables by default (supports JSON and YAML).