Skip to main content
OneX governance brings AI coding assistants-Claude Code, GitHub Copilot, and other supported tools-under the same audit and policy controls as the rest of your OneX estate. It complements model observability: where the SDK instruments models in application code, the governance client connects your IDE or CLI to OneX without code changes. Once your tenant has governance enabled, install the OneX Governance client on your workstation and register the assistants your teams already use.

​
How it works

Claude Code, Cursor, and most CLI agents connect through the governance client installer. GitHub Copilot in Visual Studio Code requires the OneX VS Code extension as well-see GitHub Copilot in VS Code.

​
Requirements

Before installation, confirm you have:
  • A OneX tenant with governance enabled
  • Your tenant ID and forwarder secret
Forwarder secret - Available in the OneX dashboard under your account.
  • The API URL for your environment, reachable from your workstation
  • The governance client package for your platform (Installation)
  • A supported IDE or CLI (listed below)
  • Git, with your project checked out in a repository (recommended for repository context)

​
Supported tools

Specify tools as a comma-separated list when you run install-client (for example claude_code,copilot).
ToolInstall nameSetup
Claude Codeclaude_codeInstallation
GitHub CopilotcopilotInstallation and GitHub Copilot in VS Code
CursorcursorInstallation
OpenAI Codex CLIcodexInstallation
WindsurfwindsurfInstallation
Gemini CLIgeminiInstallation
Antigravity IDE/CLIantigravityInstallation
KirokiroInstallation
Claude Desktop (the standalone chat application) is not supported at this time.

​
Client packages

PlatformPackageInstall command
WindowsOneXGovernance-Portable.zipOneXGovernance.exe install-client
LinuxOneXGovernance-linux-<arch>.tar.gz./OneXGovernance install-client
macOSOneXGovernance-macos-<arch>.tar.gz./OneXGovernance install-client
Your organization provides the build that matches your operating system and processor architecture. Linux and macOS each ship separate packages for x86_64 and arm64.

​
Model observability vs. agent governance

Model observability (SDK)AI agent governance
ScopeML models in production workloads (PyTorch, TensorFlow, JAX)AI assistants in the IDE and terminal
Deploymentpip install onex-sdk and OneXMonitor in your applicationGovernance client and IDE configuration
DashboardObservability dashboardGovernance dashboard in your OneX account
The two products are independent and can run on the same workstation.

​
Documentation