BrowserAct
App · AI Agents · 4 actions
Follow the BrowserAct tasks your workflows are running: check whether a task has finished, read the data its browser agent collected, and stop a run that is going nowhere.
Actions
| Action | What it does |
|---|---|
| Get task | Fetch a task by ID with everything it produced: the text result, any files, the live viewer link, and the steps the browser agent took. |
| Get task status | Check whether a task is still running. |
| List workflows | List the workflows on your BrowserAct account, newest first, with each one's ID, description, and when it was published. |
| Stop task | Stop a running task for good, useful when a workflow is taking too long or was started by mistake. |
Get task
Fetch a task by ID with everything it produced: the text result, any files, the live viewer link, and the steps the browser agent took. Run this once 'is_finished' is true to read a workflow's results.
Inputs
- Task ID
task_idtextRequired - The ID of the task to look at (the ID shown against the run on the BrowserAct task list, or the 'id' the run-task API returned).
Outputs
created_attext- When the task was created, in ISO 8601 format.
failure_reasontext- Why the task failed, or null if it did not fail.
finished_attext- When the task finished, in ISO 8601 format.
input_parameterstext- The input parameters the task ran with, as BrowserAct summarizes them (e.g. 'keyword=laptops; limit=10').
is_finishedtrue/false- True once the task is finished, failed, or canceled. False while the browser is still working; poll 'Get task status' until it turns true.
live_urltext- Link to watch the browser session live while the task runs.
output_fileslist- Links to the files the workflow produced, such as the extracted data as a spreadsheet.
output_texttext- The text result the workflow produced, or null when it produced files only or has not finished.
profile_idtext- ID of the browser profile the task used, when one was saved.
statustext- One of created, running, pausing, paused, finished, canceled, failed, or unknown.
stepslist- What the browser agent did, one entry per step.
task_idtext- The task's ID.
workflow_idtext- ID of the workflow the task ran.
Example output
{
"created_at": "2025-10-08T10:17:54Z",
"failure_reason": null,
"finished_at": "2025-10-08T10:24:09Z",
"input_parameters": "keyword=laptops; product_limit=10",
"is_finished": true,
"live_url": "https://www.browseract.com/remote/8f2c1d",
"output_files": [
"https://files.browseract.com/tasks/16429034742537847/data.csv"
],
"output_text": "Found 10 laptops between $499 and $1,299.",
"profile_id": "profile_abc123",
"status": "finished",
"steps": [
{
"evaluation_previous_goal": "No previous action taken yet; this is the initial step.",
"id": "181973",
"screenshot_url": "https://files.browseract.com/tasks/16429034742537847/step-1.png",
"status": "succeed",
"step": 1,
"step_goal": "Navigate to the search results page for 'laptops'."
},
{
"evaluation_previous_goal": "Navigation succeeded and the results page rendered.",
"id": "181978",
"screenshot_url": "https://files.browseract.com/tasks/16429034742537847/step-2.png",
"status": "succeed",
"step": 2,
"step_goal": "Extract the name and price of the first 10 products."
}
],
"task_id": "16429034742537847",
"workflow_id": "16217357109956214"
}Get task status
Check whether a task is still running. This is the cheap endpoint to poll while a task works: put it in a loop with a Wait step until 'is_finished' is true, then call 'Get task'.
Inputs
- Task ID
task_idtextRequired - The ID of the task to look at (the ID shown against the run on the BrowserAct task list, or the 'id' the run-task API returned).
Outputs
is_finishedtrue/false- True once the task is finished, failed, or canceled; false while it is still working.
statustext- One of created, running, pausing, paused, finished, canceled, failed, or unknown.
task_idtext- The task's ID.
Example output
{
"is_finished": false,
"status": "running",
"task_id": "16429034742537847"
}List workflows
List the workflows on your BrowserAct account, newest first, with each one's ID, description, and when it was published.
Inputs
- Page
pagenumber - Which page of results to return. Starts at 1.At least 1. Default:
1. - Workflows per page
limitnumber - How many workflows to return per page, from 1 to 500.From 1 to 500. Default:
50.
Outputs
limitnumber- How many workflows a page holds.
pagenumber- The page of results this is.
total_countnumber- How many workflows the account has.
total_pagesnumber- How many pages of workflows there are.
workflowslist- The workflows on this page.
Example output
{
"limit": 50,
"page": 1,
"total_count": 2,
"total_pages": 1,
"workflows": [
{
"created_at": "2025-10-07T06:15:24Z",
"description": "Collects product names and prices.",
"id": "16217357109956214",
"name": "Product price watch",
"published_at": "2025-10-07T06:27:37Z"
},
{
"created_at": "2025-10-06T11:02:10Z",
"description": "Collects company contact details.",
"id": "16217357109956215",
"name": "Lead list builder",
"published_at": "2025-10-06T11:40:55Z"
}
]
}Stop task
Stop a running task for good, useful when a workflow is taking too long or was started by mistake. A stopped task cannot be resumed or restarted.
Inputs
- Task ID
task_idtextRequired - The ID of the task to look at (the ID shown against the run on the BrowserAct task list, or the 'id' the run-task API returned).
Outputs
stoppedtrue/false- True once BrowserAct accepted the stop request.
task_idtext- The task that was stopped.
Example output
{
"stopped": true,
"task_id": "16429034742537847"
}