Runs

A run executes one script on a set of profiles. For each profile, Veilus opens the browser, runs the script against it, records the result, and closes the browser again.

Start a run

1. In Veilus Flow → Scripts, click Run on the script’s card. You can also open the script and click Test Run.
2. Under Select Profiles, tick the profiles. Use Search profiles…, Select All or Clear.
3. Under Execution Options, set:
   • Concurrency: how many profiles run at the same time (1, 2, 3, 4, 5, 8 or 10).
   • Delay Between: the pause between two profile launches (No delay, 0.5s, 1s, 2s, 3s or 5s).
4. If the script uses variables, a table shows each variable for each selected profile, filled the way the run will fill them: from the profile’s variables, its identity dataset row and the script’s defaults. A column that comes from a content dataset, or a secret dataset column, shows from dataset instead of a value. The row is picked only when the run starts, and secret values never appear in the table. Edit a cell to use a different value for this run only. The edit isn’t saved to the profile, and it takes priority over a profile variable or dataset column with the same name.
5. Click Run on N profiles.

The dialog shows a suggested maximum for Concurrency based on this computer’s current load, and warns you if you pick more.

If a required variable gets no value from the profile, its datasets or the script’s defaults, the run button stays disabled with Fill all required variables first. Fill the empty cells in the table, add the missing profile variables, or assign a dataset that has that column.

If a selected profile’s content dataset has no unused rows left, the dialog names the profile and the dataset and keeps the run button disabled, because that profile would fail as soon as the run starts. Add rows, or use Return used rows on the dataset’s page.

Only approved scripts run. Running a script on several profiles at the same time is a Pro feature, also included in the trial. If your plan doesn’t include it, set Concurrency to 1 and the profiles run one after another.

What happens during a run

For each selected profile, in order:

1. Pause between launches. Veilus waits Delay Between before starting the next profile. If CPU use is above 75% or memory use above 80%, it waits 5 seconds instead. If either is above 90%, it holds new launches until the load drops, for up to 60 seconds.
2. Wait for a turn. No more than Concurrency profiles of this run are active at once.
3. Take dataset rows. If the profile has a content dataset, its rows are reserved before the browser starts. If none are left, the profile fails here.
4. Open the browser with the profile’s fingerprint and proxy. The run shares the app-wide limit of 16 open browsers with everything else, and waits for a free slot when they are all in use. The timezone check runs here, so a proxy that now exits in a different timezone can stop the profile from launching.
5. Run the script until it exits.
6. Close the browser and record the exit code and output.

A profile fails without running the script if it is already open, if it was deleted, or if its browser can’t start. The other profiles carry on.

Read the results

Open the script and go to the Run History tab, or click History on the script’s card. Each run shows:

• where it came from: App, Schedule or Agent;
• its status: running, completed (no profile failed), failed (every profile failed), partial (some failed), or skipped;
• counts, as done · failed · total.

Click a run to see each profile’s exit code, Output (stdout) and Errors (stderr). Click Refresh to load new runs.

Run history starts from the version of Veilus that added it. Older runs aren’t linked to a script.

Runs started by an AI assistant or the API

Over MCP or the REST API:

• run_script is the trial run: up to 3 profiles, works before the script is approved, stops the script after 10 minutes.
• run_batch is the real run: the script must be approved. Concurrency defaults to 3 and goes from 1 to 16. The delay between launches defaults to 1500 ms and goes up to 10 minutes.
• get_capacity reports how many browsers are open, how many are waiting for a slot, and the maximum.
• list_runs and get_run_result read the same results the Run History tab shows.