Local installer
Run the local installer to route one machine's AI apps and agents (harnesses) through Mintive's LLM gateway. It's built for small-team installs and lightweight evaluation, so start here if you're getting a feel for how the gateway suits your needs. When you're ready to deploy to larger teams, move to MDM rollout.
The installer sets up a wide list of harnesses to route through the gateway: Claude Code, Claude Desktop, Codex, OpenCode, pi, and omp. It also manages your authentication with Mintive and your policy selection, so on first run it prompts you to sign in (OAuth or an API key), pick a policy, and choose which harnesses to configure.
macOS and Linux
curl -fsSL https://downloads.mintmcp.com/releases/mintive-cli | sh
Windows
Run this in PowerShell:
irm https://downloads.mintmcp.com/releases/mintive-cli.ps1 | iex
The mintive CLI
The installer puts a local mintive binary on your machine for managing your auth state and policy selection. After installation, run its subcommands for auth and policy operations:
| Command | What it does |
|---|---|
mintive auth status | Describe your authentication state |
mintive auth logout | Remove your local sign-in to Mintive |
mintive policy select | Interactively select your active policy |
mintive setup | Go through the full installation steps again |
See the full list with mintive --help, and add --help to any subcommand (for example, mintive policy --help) for its options.
Uninstall
When you're done trying the LLM gateway through the local installer, run mintive unconfigure and choose the harnesses to revert. It returns their configurations to how they were before the LLM gateway setup.
Next steps
- MDM rollout: deploy the gateway setup to larger teams
- Agent SDKs: point your own agents at the gateway with the token this CLI manages
- Access control: confirm which policy you have access to
- Monitoring and analysis: see your own inference calls in the Activity view