MDM installation
MDM deployment lets you roll out the AI Code Insights daemon and API credentials to every developer machine without requiring individual installs. Deploy via Jamf, Intune, Ansible, or your MDM platform of choice. For individual installs, see Installation. For rollout strategy, see Rollout.
Installer behavior
Download the latest installers from Admin → AI Code Insights. All installers are code-signed with an Atlassian certificate.
Installer: .pkg
- Installs the
aicodemetricsdbinary to/usr/local/bin/ - Creates a LaunchAgent at
~/Library/LaunchAgents/com.getdx.aicodemetrics.plist - Starts the daemon as the logged-in user
- Detects and installs hooks for supported AI coding agents
If developers have repositories in macOS protected directories (Desktop, Downloads, Documents), Full Disk Access must be granted to /usr/local/bin/aicodemetricsd. Configure this via an MDM privacy preferences policy profile (PPPC).
Installer: .msi
- Installs the
aicodemetricsdbinary and a bundled Watchman binary - Creates a scheduled task that runs the daemon as the logged-in user
- Detects and installs hooks for supported AI coding agents
Installer: .tar.gz with install script
Note: Supported on Debian and Red Hat-based distributions, amd64 only.
- Installs the
aicodemetricsdbinary to/usr/local/bin/ - Creates a systemd user service at
~/.config/systemd/user/com.getdx.aicodemetrics.service - Enables and starts the service
- Detects and installs hooks for supported AI coding agents
Configuration delivery
Deliver API credentials so developers do not need to configure anything.
macOS MDM platforms can deliver credentials via a Configuration Profile using managed preferences. Managed preferences take priority over the config file — if both are present, the Configuration Profile values win.
Preference domain: com.getdx.aicodemetrics
| Key | Type | Description |
|---|---|---|
api_url |
String | Your organization’s DX API URL |
api_key |
String | Your organization’s DX API key |
log_level |
String | Log verbosity: debug, info, warn, or error (default: info) |
Option A: MDM admin dashboard: If your MDM platform supports custom preference domains (Jamf, Mosyle, Kandji, etc.), add the keys above under the com.getdx.aicodemetrics domain.
Option B: Upload a .mobileconfig file: Use this example as a starting point. Replace each of the API credential placeholders with your organization’s values, generate unique UUIDs for the two PayloadUUID fields, and upload the file to your MDM platform.
Note: The daemon reads managed preferences only at startup. If you push updated credentials via a Configuration Profile after the daemon is already running, restart it with
aicodemetricsd restartor by runninglaunchctl unloadfollowed bylaunchctl loadon the plist.
Alternatively, deploy a config file to ~/.aicodemetrics/config.json on each machine.
Credentials can be delivered via the registry at HKLM\SOFTWARE\GetDX\AICodeMetrics using the same key names as the config file (api_url, api_key, log_level). Deploy via Intune, Group Policy, or your MDM platform. Registry values take priority over the config file.
Alternatively, deploy a config file to %LOCALAPPDATA%\aicodemetrics\config.json. Restrict write access to the config directory to prevent user tampering.
Deploy a config file to ~/.aicodemetrics/config.json via Ansible, Chef, or your configuration management tool.
Minimal config file contents:
{
"api_url": "API_URL",
"api_key": "API_TOKEN"
}
For the full settings reference, see Configuration.
Upgrades
DX cannot push daemon updates directly to developer machines. Download the updated installer from Admin → AI Code Insights and deploy via your MDM platform. The installer updates the daemon in place — no configuration changes required. See Upgrades for details.
EDR and endpoint security
The daemon’s behavior — a background process that watches file system changes, hooks into developer tools, and makes outbound HTTPS calls — can trigger behavioral alerts in EDR solutions (Palo Alto, CrowdStrike, Microsoft Defender, etc.). Pre-configure exclusions before deploying to developer machines.
See Endpoint considerations for the full list of processes to whitelist and paths to exclude per platform.