Getting started

1. Sign in

1. Open the invitation email and choose a password of at least 12 characters. The link works once and expires after 24 hours; if it has, ask your administrator to send a new one.
2. Open your room's address and sign in with your username or email address and your password.
3. If your room asks for a second factor, enter the code from your authenticator app.

Closing the browser signs you out.

2. Install the agent

Do this on the Windows computer that will run automations: the device.

1. Sign in to the room on that computer. Open This computer at the bottom of the sidebar and choose Download Mercury.
2. Open the downloaded PalatineMercury.msi. Windows asks once for administrator approval.
3. The agent installs as a Windows service that starts with the computer. Its icon, the disc, appears in the notification area.

3. Register the device

1. In the room on the same computer, open This computer again and choose Register this computer. This needs the permission to manage devices; without it, ask an administrator.
2. On the computer, click Allow in the Mercury prompt.
3. The device appears under Devices, online.
4. For unattended runs, set the device's Windows credentials: under Devices, open the device's actions menu and choose Configure Windows credentials. The room never shows the password again.

If nobody can click Allow at the computer, the room shows a command to run in a terminal opened as administrator instead.

4. Add Janus to the browser

Janus lets automations work with web pages in Chrome and Edge. The agent install usually adds it to both browsers by itself.

1. Open This computer. If Janus reads Paired, this step is done.
2. If it reads Not installed, add it from the Chrome Web Store: Palatine Janus. Edge installs it from the same page.
3. If it reads Managed browser, your company manages the browser: ask your IT administrator to add Janus.

5. Build a first task

1. Open Automations, choose + New, then New task. Type a name and press Enter. The editor opens on your draft.
2. Press / to add a step: type the step's name, for example Message box, and press Enter.
3. Fill in its fields in Properties, beside the outline.
4. For a step that acts on an application or a web page, choose Capture, point at the element on the device and click it. Use the arrow keys to pick a larger or smaller element first.

The draft saves by itself as you work. To try it at once, choose Run in the editor (or press Ctrl Enter) and pick your device.

6. Publish

1. Open Automations. On the Drafts side, select your task and choose Publish.
2. Write a short note on what changed and confirm. The task is now release v1 under Releases.

Schedules, triggers and queues run releases only, so publish each change you want them to use.

7. Run it and read the run page

1. Under Releases, select the task and choose Run. Pick the device, fill in any inputs, and confirm.
2. Open Runs and select the run.

The run page reads from the top down:

• The verdict. Done, with the time it took; or the step that failed, with its error; or, while it runs, the step it is on, with Pause and Stop beside it.
• The facts of the run in one strip.
• The timeline. Every step with its time and inputs, the log lines under the step that wrote them, beside the screen recording. Select a step and the video moves to it.
• The results. The output variables and any files the run produced.

8. Schedule it

A room schedules processes only unless an administrator changes that, so the usual way is a small process that runs your task.

1. Open Automations, choose + New, then New process. It starts with Try, Catch and Finally.
2. Inside Try, add a Run task step and pick your task.
3. Publish the process.
4. Open Schedules and click an empty slot in the calendar, or choose Schedule. Pick the process, when it runs, its environment (DEV, QA or PRD) and the device or pool, then save.

The calendar now shows the planned runs, dashed, and each run once it starts. Creating a schedule needs the permission to manage automations; administrators have it.