Bulk Run with input parameters lets you execute hundreds or thousands of different searches automatically using a single robot. Upload a CSV file with all your search terms and extract up to 50,000 searches at once.
What are input parameters?
Input parameters are variables created when you type text into form fields during robot training. They make your robot flexible - train once with one search term, then run the same robot with thousands of different terms without retraining.
Example: Train your robot searching for "laptop", and the search term becomes a parameter you can change to "tablet", "smartphone", or any other term when running the robot.
📖 You can learn more about Input Parameters and how to train a robot to extract data from a search result.
Prerequisites
Before you can perform bulk searches, you need:
✅ A trained robot with search functionality
Robot must fill a search box during training
Search term must be an input parameter
✅ A list of search terms
Needs be formatted as CSV for the upload
Maximum 50,000 rows per bulk run
Step-by-step guide
📘 This guide assumes you've already trained and approved a robot to extract data from a search query.
Step 1: Prepare your CSV file
Create a CSV with column names matching your robot's parameters:
Basic search (single parameter):
searchTerm laptop tablet smartphone wireless headphones gaming keyboard
Advanced search (multiple parameters):
searchTerm,location,minPrice,maxPrice laptop,New York,500,1500 tablet,Los Angeles,200,800 smartphone,Chicago,300,1000
💡 Column names must match parameter names exactly (case-sensitive). If you're unsure, you can download a sample CSV.
Step 2: Upload and run bulk search
Go to your robot → Click "Run Task" tab
Click "Bulk Run Tasks" button
Upload your CSV file
Review the preview to ensure parameters mapped correctly
Confirm to start the bulk run
Browse AI will perform each search and extract results
Practical examples
Example 1: E-commerce price research
Scenario: Compare prices across 100 products
Robot trains on: Product search → Extract name, price, availability
CSV file
searchTerm
iPhone 15 Pro
Samsung Galaxy S24
iPad Air MacBook Pro 14
Sony WH-1000XM5
[... 95 more products]
Result: Table with all products and their current prices
Example 2: Job market analysis
Scenario: Research salaries for roles across cities
Robot trains on: Job search with location → Extract title, salary, company
CSV file:
jobTitle,location Software Engineer,San Francisco Software Engineer,Austin Software Engineer,New York Data Scientist,San Francisco Data Scientist,Austin Product Manager,San Francisco [... more combinations]
Result: Salary data comparison across roles and locations
Example 3: Competitor monitoring
Scenario: Track product listings for multiple brands
Robot trains on: Brand search → Extract product listings
CSV file:
brandName,category
Nike,Running Shoes
Adidas,Running Shoes
Puma,Running Shoes
Nike,Basketball Shoes
Adidas,Basketball Shoes
Result: Complete inventory snapshot by brand and category
Example 4: Lead generation
Scenario: Find businesses in different cities
Robot trains on: Directory search → Extract business details
CSV file:
businessType,city,state
restaurants,Austin,TX
restaurants,Dallas,TX
dentists,Austin,TX
lawyers,Houston,TX
accountants,Austin,TX
Result: Business contact database organized by type and location
Advanced techniques
Using wildcards and variations
Include search variations to catch more results:
searchTerm
laptop
laptops
laptop
computer
notebook
computer
portable computer
💡 You can use an LLM (ChatGPT, Claude, etc.) to generate a list for you.
Combining with other parameters
You can use both Origin URL and search based on how the site you're extracting from is structured.
originUrl,searchTerm
https://site.com/electronics,laptop
https://site.com/electronics,tablet
https://site.com/furniture,desk
https://site.com/furniture,chair
Building search terms programmatically
Generate comprehensive search lists using python or LLMs.
Python example:
products = ['laptop', 'tablet', 'phone']
brands = ['Apple', 'Samsung', 'Dell']
searches = [f"{brand} {product}" for brand in brands for product in products]
# Creates: Apple laptop, Apple tablet, Samsung laptop, etc.
LLM example (ChatGPT, Claude, etc.):
``` Generate a CSV list of search terms combining these elements: - Product types: laptop, tablet, smartphone - Brands: Apple, Samsung, Dell, HP, Lenovo - Years: 2023, 2024 Format: brand product year
Monitoring and management
Track progress via robot's History tab.
Check success rate in usage reports.
Set up notifications for completion.
Review failed searches for reprocessing.
Integrations and automations with other features
Scheduling bulk searches
Combine with monitoring to check for updates for specific input parameters.
Workflows with search
Connect search results to other robots to extract data across multiple pages.
Robot A: Search and get result URLs
Robot B: Extract details from each URL
API automation
Trigger bulk searches programmatically using the API.
POST /robots/{robotId}/bulk-run { "data": [ {"searchTerm": "laptop", "location": "NYC"}, {"searchTerm": "tablet", "location": "LA"} ] }Limitations
Cannot run a bulk search with:
Dropdown selections (fixed during training)
Checkboxes or radio buttons
Image uploads
Interactive filters (sliders, maps)
Workarounds:
Create separate robots for different filter combinations
Use direct URL parameters if available
Chain multiple robots for complex searches