Browser Extension¶
Open web pages in your default browser.
Built-in
This extension is always available and requires no setup.
What You Can Do¶
Open URLs¶
In current window:
Open a link in your browser.
In new tab:
Keep your current page and open the link in a new tab.
In new window:
Open the link in a completely separate browser window.
Example requests:
- "Open the Billomat website"
- "Open this invoice link in a new tab"
- "Open the customer portal in a new window"
Workflow Examples¶
Quick reference:
- "Open the documentation for this product"
- "Show me the company website"
After finding data:
- "Open this customer's profile in Billomat"
- "View this invoice online"
Research:
- "Open Google and search for..."
- (Actually, just ask DeskAgent directly - it often knows the answer!)
What Opens Where¶
| Request | Result |
|---|---|
| "Open URL" | Current browser window |
| "Open in new tab" | New tab in existing window |
| "Open in new window" | New browser window |
URL Formats¶
DeskAgent understands various URL formats:
https://example.com- Full URLexample.com- Domain only (https added)http://localhost:8080- Local servers
Tips¶
-
Let DeskAgent find the URL: Instead of copying URLs, describe what you want. "Open the invoice in Billomat" is easier than finding and copying the URL.
-
New tab is usually best: Keeps your current work visible.
-
Links from data: After searching for a customer or invoice, ask to open it directly.
Common Issues¶
| Problem | Solution |
|---|---|
| "Can't open URL" | Check if the URL is valid |
| "Wrong browser opens" | Change your Windows default browser |
| "Page not found" | The URL may be incorrect or the site is down |
Browser Automation (Advanced)¶
Advanced Feature
Browser automation requires Chrome, Vivaldi, or Edge with remote debugging enabled.
What You Can Do¶
Control your browser programmatically for UI testing and automation:
- Fill forms automatically
- Click buttons and links
- Take screenshots
- Extract text from pages
- Execute JavaScript
Setup¶
Start Chrome with remote debugging:
Start-Process "chrome.exe" -ArgumentList "--remote-debugging-port=9222", "--user-data-dir=C:\temp\chrome-debug"
Or use the /chrome-debug command in DeskAgent.
Available Actions¶
Connection:
| Action | Description |
|---|---|
| Start browser | Launch browser with remote debugging |
| Check status | Verify browser connection |
| Connect | Connect to an already running browser |
Navigation:
| Action | Description |
|---|---|
| Navigate to URL | Go to a specific page |
| Get page info | Read current URL, title, and text |
| List tabs | See all open browser tabs |
| Switch tab | Change to a different tab |
Forms & Input:
| Action | Description |
|---|---|
| Find forms | Discover all forms on a page |
| Fill field | Enter text in a single input |
| Fill form | Fill multiple fields at once |
| Select dropdown | Choose an option from a dropdown |
| Type text | Simulate keyboard typing |
| Press key | Send Enter, Tab, Escape, etc. |
Interaction:
| Action | Description |
|---|---|
| Click element | Click by CSS selector |
| Click text | Click by visible text |
| Wait for element | Pause until element appears |
Content & Screenshots:
| Action | Description |
|---|---|
| Get text | Read text from an element |
| Execute JavaScript | Run custom JS code |
| Take screenshot | Capture the current page |
Example Workflow¶
UI Testing:
1. "Start Chrome with debugging"
2. "Navigate to http://localhost:5005"
3. "Fill the username field with 'admin'"
4. "Fill the password field with 'secret'"
5. "Click the login button"
6. "Wait for the dashboard to load"
7. "Take a screenshot"
Form Automation:
1. "Connect to the browser"
2. "Find all forms on the page"
3. "Fill the form with name 'John' and email 'john@example.com'"
4. "Click Submit"
Tips¶
- Use CSS selectors: Precise targeting with
#id,.class, or[attribute] - Wait for elements: Pages may load asynchronously
- Check connection: Make sure browser debugging is active
- Screenshots help: Verify your automation worked correctly