CLI overview

The Ramp CLI is the terminal surface for developers and AI agents. Use it when an agent can run shell commands, when you need structured output for a script, or when a packaged skill should guide a multi-step workflow.

The CLI is open source: github.com/ramp-public/ramp-cli.

Install

The install script detects your platform, downloads a pre-built binary, and sets up the ramp command:

curl -fsSL https://agents.ramp.com/install.sh | sh

If you already use uv, you can install it as a tool instead:

uv tool install ramp-cli

Authenticate

The CLI uses OAuth. Run the login command to open a browser-based flow:

ramp auth login

Confirm which user the CLI is acting as:

ramp users me

Environments

By default, the CLI connects to Sandbox. Switch per-command with --env, or set a default when you are ready to work against production:

# Per command
ramp transactions list --env production
 
# Set the default environment
ramp env production

How actions are attributed

Every CLI action runs as the user who authenticated via ramp auth login:

• Actions are attributed to you — the CLI acts as the authenticated Ramp user.
• Permissions match your role — if you can't see it in the dashboard, you can't see it in the CLI.
• Data visibility follows your access — non-admins see only their own data; admins can query across the business.

When to use MCP instead

Use MCP when the agent is a conversational assistant that already supports remote MCP. Use the CLI when the agent has terminal access, needs repeatable command output, or should follow a CLI skill like agentic-purchase.

Next: Commands & resources.