Desktop app

The jawb desktop app is a supervision console, not a browser. Your agent browses with jawb; you use the app to approve what jawbd holds for a person, review what the agent did, and unblock it. It is a client of jawbd like the CLI: jawbd runs every action and applies every policy, and the app holds no state of its own.

Download it from the download page, which offers the build for your system first and lists every platform: macOS (Apple silicon and Intel), Windows x86_64 and Linux x86_64. The first desktop release is not published yet; the download links work once it is.

What it shows

• Approvals. Every action jawbd holds for a person, each on its own card: a purchase with the site, the amount and who asked, a change that weakens a safety setting, an extension install. You approve or deny it there, typing the word the card asks for, such as purchase.
• Activity. The trace of what your agent did, newest first, the same steps as jawb trace.
• Extensions. What is installed, where it came from, and what each extension may do.
• Status. Whether jawbd is running and which version.

The app reads from jawbd when you open a view, when its window gains focus, and when you click Refresh. It runs no timers, so an idle window sends nothing. A hold that arrives while you are not looking is announced by jawbd itself, through the other approval channels.

The daemon

The app ships with its own jawbd. When no daemon is running, it starts that one; when the CLI already started one, it connects to it. It never stops a jawbd it did not start.

Opening it the first time

The desktop builds are not code-signed yet, so macOS and Windows ask you to confirm the first launch:

• macOS. Open the .dmg and drag jawb to Applications. Gatekeeper blocks the first launch: open System Settings → Privacy & Security, and click Open Anyway next to the message about jawb.
• Windows. Run the setup .exe. SmartScreen says "Windows protected your PC": click More info, then Run anyway.
• Linux. Make the AppImage executable (chmod +x), then run it.

Requirements

Nothing beyond the app itself for the console. Act and See commands your agent runs still need a locally installed Chrome or Chromium, as in the installation guide.

Next: run the daemon.