Documentation Index

Fetch the complete documentation index at: https://docs.trynavi.guide/llms.txt

Use this file to discover all available pages before exploring further.

​

Changelog

​

[1.6.0] - 2025-09-25

​

Added

• Company Tracking: New company parameter for enhanced user tracking and analytics
  • Track users by company or organization name
  • Filter conversations by company in the dashboard
  • Generate company-specific reports and usage statistics
  • Supports both static and dynamic company assignment
  • Integrates with existing user tracking for comprehensive analytics

​

Usage

// Basic company tracking
initUnifiedWidget({
  agentConfigId: "your-agent-id",
  agentKey: "your-agent-key",
  user_identifier: "user-123",
  company: "Acme Corporation" // Track by company
});

// Dynamic company assignment
initUnifiedWidget({
  agentConfigId: "your-agent-id",
  agentKey: "your-agent-key",
  user_identifier: String(currentUser.id),
  company: currentUser.organization // Use from user data
});

​

[1.5.0] - 2025-08-21

​

Added

• Widget Initialization Callback: New onInitialized callback that provides precise control over widget initialization state
  • Called when initialization completes (success or failure)
  • Provides success boolean, error message (on failure), and widgetReady status
  • Enables developers to handle initialization states properly
  • Supports both authorization and configuration error scenarios

​

Usage

// Handle widget initialization state
initUnifiedWidget({
  agentConfigId: "your-agent-id",
  agentKey: "your-agent-key",
  onInitialized: (result) => {
    if (result.success) {
      console.log("Widget ready!");
      // Enable UI elements, show controls, etc.
    } else {
      console.error("Widget failed:", result.error);
      // Show error state, disable features, etc.
    }
  }
});

​

[1.4.0] - 2025-08-20

​

Added

• Memory Context for Onboarding Agents: New enableMemoryContext option that enables intelligent conversation continuity for onboarding agents
  • Automatically tracks user progress across conversations
  • Stores Q&A responses and milestones completed
  • Personalizes agent responses based on previous interactions
  • Dynamically overrides first message for returning users
  • Only works with agents of type “onboarding”

​

Usage

// Enable memory context for onboarding continuity
initUnifiedWidget({
  agentConfigId: "your-onboarding-agent-id",
  agentKey: "your-agent-key",
  user_identifier: "user123", // Required for memory context
  enableMemoryContext: true // Enable intelligent conversation continuity
});

​

[1.3.0] - 2025-08-19

​

Added

• Multilingual Voice Support: Added support for 32 languages in voice conversations
  • Configure default language using defaultLanguage.voice parameter
  • Supports all major languages including Spanish, French, German, Chinese, Japanese, and more
  • Language is fixed for the duration of the conversation
  • Works with ElevenLabs provider using v2.5 Multilingual model

​

Changed

• Panel Configuration Structure: Moved title and description into panelConfig object
  • Old: title and description as top-level properties
  • New: panelConfig: { title, description, expanded }
  • Provides better organization and adds expanded option

​

Usage

// Language support
initUnifiedWidget({
  agentConfigId: "your-agent-id",
  agentKey: "your-agent-key",
  defaultLanguage: {
    voice: "es" // Spanish
  }
});

// Panel configuration
initUnifiedWidget({
  agentConfigId: "your-agent-id",
  agentKey: "your-agent-key",
  name: "Assistant",
  panelConfig: {
    title: "Welcome",
    description: "How can I help you today?",
    expanded: false
  }
});

​

[1.2.4] - 2025-08-18

​

Changed

• Usage Limit Checking Now Enabled by Default: Usage checking is now automatically enabled when user_identifier is provided
  • Previously required explicitly setting enableUsageChecking: true
  • Now checks usage limits automatically to help manage costs
  • Can be disabled by setting enableUsageChecking: false
  • Fail-safe behavior: If usage check fails, widget still initializes

​

Usage

// Usage checking is now automatic when user_identifier is provided
initUnifiedWidget({
  agentConfigId: "your-agent-id",
  agentKey: "your-agent-key",
  user_identifier: "user123", // Automatically triggers usage checking
});

// To disable usage checking (not recommended for production)
initUnifiedWidget({
  agentConfigId: "your-agent-id",
  agentKey: "your-agent-key",
  user_identifier: "user123",
  enableUsageChecking: false, // Explicitly disable
});

​

[1.2.1] - 2025-08-14

​

Added

• Widget Customization: New properties to customize the widget’s appearance and messaging
  • name: Custom name shown in the widget header
  • title: Title displayed in the expanded panel
  • description: Description text explaining the assistant’s capabilities
  • Perfect for branding and tailoring the experience to different use cases

​

Improved

• Voice UI Enhancements: Updated the voice interface with improved styling and user experience

​

Usage

initUnifiedWidget({
  agentConfigId: "your-agent-id",
  agentKey: "your-agent-key",
  
  // Widget Customization
  name: "Shop Assistant",
  title: "Need Help Shopping?",
  description: "I'm your personal shopping assistant! I can help you find products, compare prices, and guide you through checkout.",
});

​

[1.1.0] - 2025-08-12

​

Added

• Conversation Persistence for Voice: New enableConversationRestore option that allows voice conversations to continue seamlessly after page refreshes
  • Conversations automatically restore if refreshed within 5 minutes
  • The AI assistant remembers the context and continues naturally
  • No need to repeat information after navigating between pages
  • Enable with enableConversationRestore: true in your configuration

​

Usage

initUnifiedWidget({
  agentConfigId: "your-agent-id",
  agentKey: "your-agent-key",
  enableConversationRestore: true, // Enable conversation persistence
});

​

[1.0.9] - Previous Release

• Base functionality for chat and voice interactions
• Multi-provider support
• Tool execution framework