You can train a robot to extract data behind a login using two methods:
Using your session cookies (recommended for most cases)
Using your user credentials (username and password)
Quick decision guide
If your site has... | Use this method | Why |
Standard login (no 2FA) | Session cookies | Fewer steps, higher success rate |
Two-factor authentication | Session cookies | Bypasses 2FA complexity |
IP-restricted access | Username/password | Cookies may be rejected from different IPs |
Frequently changing login page | Session cookies | Avoids retraining for UI changes |
High-security requirements | Username/password | Some sites block external cookies |
Method 1: Login with session cookies (recommended)
Session cookies store your authenticated state. When you're logged into a website, the robot can use these cookies to access protected content without going through the login process.
How to set up cookie-based login
β You must be logged into the website in your browser before starting.
Start training your robot in Robot Studio.
Select This website needs logging in, and then select Login with my session cookies.
Your cookies are then automatically safely encrypted and stored securely. From here continue to train your robot to extract and monitor the data you need.
Benefits of session cookies
No login steps during extraction (faster, more reliable)
Works with most 2FA/MFA systems
No need to record login interactions
Higher success rates overall
Limitations of cookies
Cookies expire (need periodic updates)
Some sites reject cookies from different IPs
Not guaranteed with all 2FA implementations
Maintaining cookie-based login
When to update cookies:
Robot starts failing with authentication errors
After you've logged out and back in
Following password changes
Periodic refresh (monthly recommended)
How to update:
β You must be logged into the website in your browser first.
Go to your robot β Settings tab
Find Authentication section
Click Update Session Cookies
Method 2: Login with username and password
Train your robot to enter your username and password and submit them, mimicking human login behavior.
How to set up credential-based login
Start training your robot in Robot Studio.
Select This website needs logging in, and then select Login with my password.
Login using your username and password during training and click to login.
π‘ Your credentials are encrypted and stored on Browse AI's secure AWS infrastructure. We are SOC II Type 2 certified, you can learn more about our security measures here.
Benefits of using credentials
Works on IP-restricted sites
Doesn't expire like cookies
Consistent across different locations
No manual updates needed
Limitations of credentials
Additional interactions (typing, clicking) introduce more potential failure points
May not work well with websites that frequently change their login interface
Can be problematic with sites that use 2FA/MFA
May trigger security alerts if the login occurs from different IP addresses
Editing or updating login credentials
If your password or login details have changed, or if you're experiencing issues with credential login you'll need to retrain your robot.
Go to your robot β Settings tab
Find Danger Zone section
Click Re-train robot
π‘ When training your robot with login credentials it's best to keep the login process as simple as possible.
Choosing the right method
Use session cookies when:
Site has 2FA/MFA authentication
You want maximum reliability
Login page changes frequently
You're already logged in regularly
Speed is important
Use credentials when:
Site restricts cookie usage by IP
You need fully automated solution
Cookies consistently fail
Site has simple login (no 2FA)
Running from multiple locations
Common authentication types
Auth type | Cookie method | Credential method | Notes |
Basic login | β Works well | β Works well | Either method suitable |
2FA (SMS/App) | β Best choice | β Usually fails | Cookies bypass 2FA |
SSO (Google/Microsoft) | β Recommended | β οΈ Complex | Multiple redirects problematic |
CAPTCHA on login | β Bypasses | β οΈ May work | See CAPTCHA handling |
Biometric | β After auth | β Not possible | Use cookies post-authentication |
Security considerations
How Browse AI protects your data
All login information (both cookies and credentials) is:
Encrypted during transmission and storage
Stored on secure AWS infrastructure
Only used for the purpose of running your robots
Website detection risks
While Browse AI employs techniques to mimic human browsing, some websites with strict security may detect automated logins, especially when:
Logins occur from multiple IP addresses
Multiple requests are made in short succession
Interaction patterns differ from typical human behavior
If your automation needs pose this risk, consider local automation, which utilizes your personal IP address, for enhanced discretion. For public data extraction, our robots are a reliable and secure solution.
Troubleshooting login issues
Diagnostic checklist
Issue | Cookie method fix | Credential method fix |
"Not logged in" error | Update session cookies | Verify credentials still valid |
Partial data extraction | Check cookie permissions | Login may be incomplete |
Intermittent failures | Cookies may be expiring | Site may have rate limiting |
Sudden stop working | Site may have logged you out | Login page likely changed |
2FA challenges appearing | Refresh cookies after 2FA | Not fixable - switch to cookies |
Common solutions
Verify manual login works
Test credentials manually
Check for new security features
Note any new steps required
For cookie failures:
Clear browser cache
Log out and back in
Update cookies immediately
Try incognito mode
For credential failures:
Retrain robot with current flow
Simplify login steps
Check for CAPTCHA
Consider switching to cookies
When authentication won't work
Some sites cannot be accessed:
Hardware token requirements
Biometric-only authentication
IP whitelist restrictions
Advanced bot detection
Alternatives:
Check for API access
Export features on the site
Manual extraction
Browse AI managed services
Best practices
Do's
β
Test login method before scaling
β
Monitor success rates regularly
β
Update cookies proactively
β
Use the simplest method that works
β
Document which method you chose
Don'ts
β Share login credentials
β Use compromised accounts
β Violate terms of service
β Ignore security warnings
β Extract sensitive personal data




