One command: opencues update. It runs git pull and then re-runs the installer for every integration it detects on your machine. The docs call it the 'I don't want to think about it' command, and that's the intent.
What update covers, and what it deliberately doesn't
opencues update rebuilds and redeploys host integrations. It deliberately does not rewrite your ~/.cues/ config tree; pulling new shipped cues and blanks onto disk is a separate, narrower command, opencues update-configs, so that rebuilding a host can never silently touch your configs. Your own edits are preserved either way: seeding only copies files you don't already have, and library-script syncing never overwrites .md content.
The safety net: self-healing launches
Even if you forget to update after a pull, the launch path covers you. Every opencues run <host> compares a hash of the current runtime source against the hash recorded when that host's bundle was installed. If any source byte changed, it rebuilds before launching, with a one-line notice (bundle is stale ... Rebuilding before launch); pass --no-rebuild-check to skip. This closes the classic trap where a patched host silently runs weeks-old code while the repo looks current. Direct binary launches bypass the check, which is one reason opencues run is the recommended path.
Verifying afterwards
opencues version: prints the CLI and every per-integration version, with compatibility markers.opencues doctor: flags any install the update pass missed, including per-install drift warnings.
Chrome is the one host with an extra step by nature: after a rebuild, the extension needs a reload in chrome://extensions, and config changes reach it via opencues sync chrome.
Related: opencues run · How to install · opencues doctor · Release notes · GitHub repo