Tab Selection

The /tab: tag provides an intelligent way to include browser tab context in your AI conversations. Select one or multiple tabs from Chrome or Safari to give the AI information about what you're working with.

Quick Start

Type /tab: in the launcher to open the tab selector:

/tab: Summarize these articles
/tab: Compare these documentation pages
/tab: What's the relationship between these tabs?

Tab Selector Interface

Visual Elements

• Search Bar - Filter tabs by title, domain, or URL
• Tab List - All open tabs from Chrome and Safari
• Checkboxes - Visual indicators for selected tabs
• Selection Count - Shows number of selected tabs in footer
• Browser Icons - Chrome (globe) or Safari icon for each tab
• Active Indicator - Star shows currently active tab

Selection Methods

Single Selection

Click any tab to select it and immediately confirm. Perfect for quick, single-tab context.

Multi-Selection with Keyboard

• Cmd+Click - Toggle individual tab selection
• Shift+Click - Select a range of tabs
• Cmd+A - Select all visible (filtered) tabs
• Space - Toggle selection of highlighted tab

Navigation

• ↑/↓ Arrow Keys - Navigate through tab list
• Enter - Confirm current selection
• Escape - Cancel and close selector

Multi-Tab Selection Examples

Research Workflow

1. Open multiple research articles in browser tabs
2. Type: /tab: [Select all research tabs]
3. Ask: "Summarize the key findings from these papers"

Documentation Comparison

1. Have several documentation pages open
2. Type: /tab: [Select relevant docs]
3. Ask: "What are the differences between these API versions?"

Code Review

1. Open multiple code files in browser tabs
2. Type: /tab: [Select code tabs]
3. Ask: "Find potential issues across these files"

Meeting Preparation

1. Have calendar, notes, and project pages open
2. Type: /tab: [Select all relevant tabs]
3. Ask: "Create an agenda based on these resources"

Advanced Features

Smart Filtering

The search bar supports multiple search strategies:

• Domain search: Type "github" to see all GitHub tabs
• Title search: Type keywords from page titles
• URL search: Type partial URLs to find specific pages

Selection Persistence

Your selections are preserved even as the tab list updates:

• Tab list refreshes every 5 seconds
• Selected tabs remain selected if they still exist
• Closed tabs are automatically removed from selection

Intelligent Updates

The selector only refreshes when tabs actually change:

• Reduces visual disruption
• Maintains scroll position
• Preserves your workflow

Tab Information Provided

When you select tabs, the AI receives:

Single Tab

[Selected Tab Information]
URL: https://example.com/page
Title: Page Title
Browser: Chrome
Domain: example.com

Note: User has selected this tab from their open browser tabs.

Multiple Tabs

[Selected Tabs Information - 3 tabs]

Tab 1:
- URL: https://example.com/page1
- Title: First Page
- Browser: Chrome
- Domain: example.com

Tab 2:
- URL: https://docs.example.com/api
- Title: API Documentation
- Browser: Safari
- Domain: docs.example.com

Tab 3:
- URL: https://github.com/repo
- Title: Repository
- Browser: Chrome
- Domain: github.com

Combining with Other Features

With Agent Mode

/agent /tab: [Select project tabs] "Create a status report"

With Vision Mode

/tab: [Select design tabs] /screen "Compare these designs with what's on screen"

With Model Selection

/model:genius /tab: [Select complex docs] "Explain the architecture"

Tips and Best Practices

Efficient Selection

1. Use search first - Filter before selecting to work faster
2. Keyboard shortcuts - Cmd+A after filtering selects visible tabs
3. Range selection - Shift+Click for consecutive tabs
4. Clear search - X button or Escape clears filter

Workflow Optimization

1. Group related tabs - Open related content in adjacent tabs
2. Use browser windows - Organize projects in separate windows
3. Name tabs meaningfully - Helps with searching and selection
4. Close unnecessary tabs - Reduces clutter in selector

Privacy Considerations

1. Selective sharing - Only selected tabs are shared with AI
2. URL only - Full page content requires explicit extraction
3. No automatic access - You control what tabs are included
4. Clear indicators - Selected tabs are visually marked

Troubleshooting

Tabs Not Appearing

• Ensure Chrome or Safari is running
• Check browser permissions in System Settings
• Restart Agentastic if tabs aren't detected
• Some browser profiles may not be accessible

Selection Issues

• Click checkbox or use Cmd+Click for multi-selection
• Single click without modifier selects only that tab
• Space key toggles current highlighted tab
• Escape cancels selection entirely

Performance

• Large numbers of tabs may slow initial load
• Filtering improves performance with many tabs
• Consider closing unused tabs for better experience
• Tab refresh interval is optimized at 5 seconds

Keyboard Shortcut Reference

Integration with Browser Tools

The tab selector works seamlessly with Agentastic's browser tools:

1. Content Extraction - AI can extract full page content when needed
2. Screenshots - Capture visual content from selected tabs
3. Navigation - AI can navigate to specific tabs
4. Interaction - Perform actions on selected tabs

Privacy and Security

What's Shared

• Only tab metadata (URL, title, browser)
• No automatic content extraction
• No cookies or session data
• No private browsing tabs

User Control

• Explicit selection required
• Visual confirmation of selected tabs
• Can cancel at any time
• Clear indication of what will be shared

Next Steps

• Learn about Browser Tools for content extraction
• Explore Screen Capture for visual context
• Master Agent Mode for complex browser automation

The tab selector makes it easy to give AI context about your browser workspace, enabling more relevant and helpful responses!