Publishing Your Agent

Complete guide to publishing and sharing your AI agent across different platforms

Publishing Your AI Agent

The publishing process is the final step in making your AI agent available to users. This involves configuring all settings, testing functionality, and generating sharing links or embed codes.

Publishing Workflow Overview

Based on the screenshots from the sharing interface, the publishing process follows these key steps:

Step 1: Agent Configuration

The main customization interface shows the core configuration areas:

• Face Focus vs Body Focus: Choose between close-up or full-body avatar display
• Avatar Selection: Choose from available avatars or upload custom ones
• Companion Name Configuration: Set identity (Sales1, Marketing, Support1, etc.)
• Agent Name: Define the agent’s display name (e.g., “Product Presenter”)
• Voice Selection: Choose Gemini voices (Sportsman, Customer support, Sarah, Brooke, Katie, Zemo, ajith, duaila, azj, ajz, sjl, brit, Swissen) or OpenAI Realtime voices (Alloy, Echo, Shimmer, Ash, Ballad, Coral, Sage, Verse, Cedar, Marin) depending on the active model
• LLM Model: Select from OpenAI GPT 4.1/5 series, Gemini 2.5 series, OpenAI Realtime family, or Groq-hosted models

Step 2: Behavior Setup

Configure your agent’s conversational behavior:

• Agent Prompt: System instructions that guide responses
• Initial Agent Messages: Welcome message (e.g., ”👋 Hi, how can I help you?”)
• Initial Suggested Messages: Pre-written prompts to help users start conversations
  • “Give me a summary of this knowledge base”
  • “Can you provide an overview of the main topics covered here?”
  • “What categories or sections does this knowledge base cover?”

Step 3: Advanced Configuration

Set up advanced features:

• MCP Agent Integration: Select and configure MCP (Model Context Protocol) agents
• Knowledge Base Integration: Connect relevant knowledge sources
• Companion Voice Settings: Fine-tune voice characteristics
• LLM Modal Selection: Choose appropriate AI models for your use case

Step 4: Voice and Realtime Settings

Configure voice and realtime capabilities:

• Realtime Model Families: Select from the latest offerings:
  • OpenAI Realtime: GPT Realtime, GPT‑4o Realtime, GPT Realtime Mini
  • Gemini 2.5: Flash Lite, Flash, Pro
  • Groq hosted: GPT OSS 20B, GPT OSS 120B, Qwen3‑32B, Moonshotai Kimi K2
• Voice Options: Choose from multiple voice personalities
• Real-time Processing: Enable live conversation capabilities

Step 5: Model Selection Details

The interface shows expanded model options:

• OpenAI GPT series: GPT 4.1 mini, GPT 4.1, GPT 5, GPT 5 nano, GPT 5 mini
• Gemini Models: Gemini 2.5 Flash Lite, Gemini 2.5 Pro, Gemini 2.5 Flash
• OpenAI Realtime: GPT Realtime, GPT‑4o Realtime, GPT Realtime Mini
• Groq hosted: GPT OSS 20B, GPT OSS 120B, Qwen3‑32B, Moonshotai Kimi K2
• Creative Stack: Google Nano Banana, Veo 3.1, Sora 2 (auto-selected for image/video generation prompts)
• Model-specific Features: Each family balances latency, reasoning depth, and integration compatibility differently

Experience Customization

Step 6: Visual Experience Setup

Configure the visual experience:

• Companion Background Color: Choose from color palette or custom hex codes
• Scene Selection: Pick from multiple visual environments:
  • Office settings
  • Presentation rooms
  • Tech environments
  • Custom backgrounds
• Logo Upload: Add company branding (1600x30 pixels recommended)

Step 7: Avatar Customization

Select and customize your agent’s appearance:

• Pre-built Avatars: Professional business avatars available
• Diverse Options: Multiple ethnicities and styles
• Custom Upload: Upload your own avatar images
• Voice Matching: Select voices that match the avatar personality

Step 8: Theme Confirmation

The system shows confirmation when theme updates are successful:

• Agent Preview Theme Updated: Visual feedback confirming changes
• Real-time Preview: See changes immediately in the interface
• Consistency Check: Ensure all visual elements work together

Data Collection Setup

Step 9: Lead Capture Forms

Configure data collection:

• Form Title: Set descriptive titles (e.g., “Contact details”)
• User Consent Text: Add privacy-compliant language
• Field Configuration:
  • Name fields with custom placeholders
  • Email fields for follow-up
  • Mandatory vs optional fields
• Form Workflow: “Add new input”, “Back”, “Confirm” options

Step 10: Multi-field Forms

Set up complex data collection:

• Multiple Field Types: Name, Email, custom fields
• Field Validation: Ensure data quality
• Progressive Forms: Collect information step-by-step
• User Experience: Smooth form completion process

Publishing and Sharing

Step 11: Distribution Methods

Choose your sharing method:

• Direct Link: Simple URL sharing
• iFrame Code: Website embedding
• Chat Bubble: Floating website integration

The interface provides:

• Direct link URLs for immediate sharing
• Scene-specific URLs for customized experiences
• Social sharing buttons (WhatsApp, Email, LinkedIn, Facebook, Twitter)

Step 12: iFrame Integration

Generate embed code for websites:

<iframe src="https://chat.aivah.ai/companion/9d9d5abe-7272-4792-8e3f-ee4118b38303?type=face-focus&frame=true" 
        allow="clipboard-write; camera *; microphone *; display-capture *; video *;" 
        width="100%" height="700" frameborder="0">
</iframe>

Step 13: Bubble Script Integration

JavaScript code for chat bubble integration:

<script>
window.aivahUrl = "https://models.readyplayer.me/68971051680864d7c650276d.png"
window.aivahBackground = "#373737"
window.baseUrl = "https://chat.aivah.ai/base?avatarId=649"
window.contentUrl = "https://chat.aivah.ai/companion/c6b3dfc8-b601-4b06-8f4a-0904a0019e3a?type=face-focus&viewType=bubble"
</script>

Agent Management

Step 14: Shared Agents Overview

Manage multiple published agents:

• Agent Gallery: View all shared agents
• Distribution Types: Custom, Chat Bubble, iframe, Direct Link
• Agent Profiles: Each showing agent name, avatar, and LLM model
• Pagination: Handle large numbers of agents (12 per page)

Step 15: Quick Sharing Access

Quick access to sharing information:

• Direct Link URLs: Immediate access links
• Scene-specific Links: Customized experience URLs
• Easy Copy: One-click copying of sharing links
• Multiple Formats: Links for different use cases

Publishing Checklist

Before publishing your agent, ensure:

Configuration Complete

• ✅ Agent behavior and prompts configured
• ✅ Avatar and visual theme selected
• ✅ Voice and model settings confirmed
• ✅ Data collection forms set up (if needed)

Testing Complete

• ✅ Test conversations in preview mode
• ✅ Verify voice functionality works
• ✅ Check mobile responsiveness
• ✅ Test data collection forms

Distribution Ready

• ✅ Choose appropriate sharing method
• ✅ Generate and test sharing links/code
• ✅ Prepare marketing materials
• ✅ Set up analytics tracking

Post-Publishing

• ✅ Monitor agent performance
• ✅ Collect user feedback
• ✅ Update configuration as needed
• ✅ Scale based on usage patterns

Best Practices for Publishing

1. Start with Testing: Always test thoroughly before publishing
2. User-Centric Design: Focus on user experience and ease of interaction
3. Clear Purpose: Ensure users understand what your agent can help with
4. Brand Consistency: Maintain consistent visual and voice branding
5. Performance Monitoring: Track usage and optimize based on data
6. Iterative Improvement: Continuously refine based on user feedback
7. Security Awareness: Consider privacy and security implications
8. Documentation: Keep records of configurations for future reference

The publishing process transforms your configured AI agent into a live, accessible tool that can serve your users across multiple channels and platforms.