DOCUMENTATION

Set up, build and run.

Practical instructions for the current desktop application, without requiring a specific platform workflow.

Getting started

Install PBX Advanced Automation on a supported Windows computer, launch the desktop app, sign in if prompted, and open your Script Library. Select a script or create a new one if your plan permits.

  1. Choose PBX Browser, Chrome or Cent Browser in the main window.
  2. Open Browser Config and select the appropriate executable or PBX API profile.
  3. Set up Proxy Config if required for an authorized workflow.
  4. Configure your script actions and optional per-script browser profile.
  5. Start Automation and inspect the result log.

Browser configuration

PBX Browser: Uses the PBX Local API and managed environments. Scripts can specify an exact environment ID. With none specified, the application can select an available stopped profile, or create one if needed.

Google Chrome and Cent Browser: Use local executables and native CDP. A script can specify a user-data directory. When none is specified, the app uses a temporary isolated profile.

Never point two running browser processes at the same user-data directory. Use a dedicated automation directory and keep a backup of important browser profiles.

Script editor and run count

The Script Editor includes an action library, editable arguments, input variables, script import/export, and per-script settings. Repeat entire script controls full workflow runs. Individual loop actions repeat only their enclosed steps.

The current runner uses one browser worker. Multiple runs execute sequentially, with fresh variables each time. Stopping or restarting does not resume an interrupted workflow.

Variables, JSON & math

Use set_variable to store text, numbers, JSON or expressions; use calculate for computed values. References such as ${server_response.user.name} and ${server_response.items.0.id} work directly.

For string merging and math, start Value with =. See the full variables and expression guide for examples and type rules.

Action categories

Actions include navigation and selectors; waiting and assertions; files and documents; iframe and drag-drop operations; HTTP/JSON; Excel/CSV and web tables; list and dynamic records; time sequence utilities; state-aware form controls; and manual approval checkpoints.

For dynamic records, use runtime record-loop actions after CSV, Excel or API parsing. The basic runtime loop has bounded limits and does not support all nested control-flow combinations.

Complete Action Reference

Find each available command by name, see its Target and Value guidance, and search across descriptions in the All Actions documentation.

Results and verification

Review the execution log for successful steps, errors, stop events and the final summary. Treat on-page success messages as preliminary when the workflow changes external business data. For payments, invoices, or documents, verify outcomes using authoritative provider records where available.

A script that starts fresh may repeat external actions. Use unique task IDs and server-side idempotency for operations that must not be duplicated.

Troubleshooting

Chrome or Cent Browser doesn't start

Confirm the configured executable path, directory permissions and that no other browser process is using the same user-data directory.

PBX environment list is empty

Check the PBX Local API connection and capabilities, then refresh the profile list. Confirm the environments exist and that the API account has access to them.

Why can't I open two browser workers?

The current application intentionally limits browser execution to one worker. Repeated script runs execute sequentially.

Why does Start begin from the first step?

This is the intended fresh-start behavior. Stop/restart does not resume an interrupted workflow.