Appearance
Usage
Browse AI has two modes of use: the no-code dashboard for training robots, and the REST API for triggering runs programmatically.
Core concepts
| Term | Meaning |
|---|---|
| Robot | A saved automation: a trained sequence of browser actions + data extraction rules |
| Task | A single execution of a robot; stores results and extracted data |
| Monitor | A recurring schedule attached to a robot (run every N hours/days) |
| Bulk Run | Execute one robot against up to 50,000 different input URLs simultaneously |
| Webhook | An HTTP callback fired after a task or bulk run completes |
No-code: train a robot in the dashboard
- Log in at https://dashboard.browse.ai/
- Click New Robot and enter the target URL
- The Chrome extension opens the target page in a recording session
- Click through the page to define actions (navigate, click, extract table, pagination, etc.)
- Browse AI's AI layer automatically identifies and structures the data fields
- Save the robot; it gets a unique
robotId - Run it manually, schedule it as a Monitor, or trigger it via API
API: run a robot task
Replace YOUR_API_KEY and YOUR_ROBOT_ID with real values.
Trigger a task (cURL)
bash
curl -X POST "https://api.browse.ai/v2/robots/YOUR_ROBOT_ID/tasks" \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"inputParameters": {
"originUrl": "https://example.com/products"
}
}'Response (201 Created):
json
{
"statusCode": 201,
"messageCode": "success",
"task": {
"id": "task_abc123",
"robotId": "YOUR_ROBOT_ID",
"status": "pending",
"createdAt": 1719700000000
}
}Poll task status and retrieve results
bash
curl "https://api.browse.ai/v2/robots/YOUR_ROBOT_ID/tasks/TASK_ID" \
-H "Authorization: Bearer YOUR_API_KEY"Response when complete:
json
{
"statusCode": 200,
"messageCode": "success",
"task": {
"id": "task_abc123",
"status": "successful",
"capturedDataItems": [
{ "name": "Price", "value": "$29.99" },
{ "name": "Title", "value": "Widget Pro" }
],
"capturedScreenshots": {}
}
}List all robots (verify API key works)
bash
curl "https://api.browse.ai/v2/robots" \
-H "Authorization: Bearer YOUR_API_KEY"Node.js example (using fetch)
js
const BROWSE_AI_KEY = process.env.BROWSE_AI_KEY;
const ROBOT_ID = process.env.BROWSE_AI_ROBOT_ID;
async function runTask(inputUrl) {
const res = await fetch(
`https://api.browse.ai/v2/robots/${ROBOT_ID}/tasks`,
{
method: 'POST',
headers: {
'Authorization': `Bearer ${BROWSE_AI_KEY}`,
'Content-Type': 'application/json',
},
body: JSON.stringify({
inputParameters: { originUrl: inputUrl },
}),
}
);
const data = await res.json();
console.log('Task created:', data.task.id, '| Status:', data.task.status);
return data.task.id;
}
runTask('https://example.com/page-to-scrape');Prebuilt robots (250+ templates)
Browse AI ships 250+ ready-to-use robots for common sites (LinkedIn, Amazon, Google Maps, Airbnb, etc.). Clone one from the dashboard under Robot Templates and use it immediately without training.
Integrations
Export results automatically to:
- Google Sheets (native integration)
- Airtable
- CSV / JSON download
- Webhooks (POST to any endpoint on task completion)
- Zapier, Make, and 7,000+ other tools via integration marketplace