Logo INTERACTIVE SANDBOX
Back to Website
B

Booking_Bot_v2

AI Agent Active

Flow Knowledge Dashboard

Agent Configuration

Used to anchor dynamic variables like {{system__time}} and calendar bookings.
LLM Temperature (0.7) Higher values make responses more creative. Lower values make them more deterministic.
Speech Speed (1.0x) Adjust the speaking rate of the text-to-speech engine.
Interruption Sensitivity How easily the agent stops speaking when it hears background noise or the user talking.

Backup LLM Redundancy

    Record Call Audio Save the raw audio files of the conversations. Disabling this saves only text transcripts.
    PII / PCI Redaction Automatically scrub credit cards, social security numbers, and sensitive info from transcripts.
    Zero Retention Mode Do not save ANY logs, transcripts, or data once the call has concluded.

    Active MCP Clusters

    Connect external APIs, microservices, and databases natively to the LLM agent via MCP.

    Inventory_API_Server
    Status: Active (Synced 18 Tools)

    Variable Data Points

    Identifier Type Description Action

    Evaluation Criteria

    Criteria Name Evaluation Instructions Action

    Document Library

    Available Library Items

    HubSpot CRM

    Expose contact properties, deal pipelines, and note logging functions to flow nodes.

    Email Templates

    Template Name Action

    Edit Action

    Add Action

    Add Audio Tag

    Explain when the agent should use this audio tag to influence the speech generated by the v3 model.

    Edit Agent Avatar

    Clone Agent

    Important: When cloning to a different account, Knowledge Base documents and integrations (Email/Calendar/MCP) are account-specific and will be stripped. Note: Default voices will clone perfectly, but Custom Cloned Voices will be removed and require re-selection.

    Welcome to the Agent Sandbox

    This is a fully interactive sandbox! You can build, configure, and connect nodes just like our live platform.

    Things to try:
    • Double-click any node to open its rich configuration editor.
    • Click AI Wizard in the toolbar to auto-generate entire conversational flows.
    • Open the Start Node to explore multi-lingual capabilities and voice cloning.
    • Click Add Action to explore the tool palette (CRM, Email, SMS, Calendar, etc.).

    Note: Since this is a static sandbox, backend functions like saving, real phone calls, or live LLM processing are simulated locally.

    AI Flow Wizard

    Describe what you want your AI agent to do, or select a quick-start template below. The wizard will automatically generate the required nodes and connections.

    Or use a Quick-Start Template:
    Dental Clinic Verifies patient and checks calendar availability.
    Restaurant Handles table bookings and takeaway orders.
    Emergency Plumber Routes emergencies via SMS and transfers calls.
    Real Estate Captures lead details and checks CRM.
    Tech Support Troubleshoots and creates API tickets.
    HR Recruiter Conducts interviews and emails summaries.

    Clone Voice

    Best Practices for a Perfect Clone
    • Length: 1 to 2 minutes is ideal. More than 3 minutes does not improve quality.
    • Quality: Ensure absolutely NO background noise, echo, or music.
    • Performance: Speak exactly how you want the AI to sound. The AI will copy your speed, emotion, inflections, and tone. Keep it consistent.

    Talk for at least 1 minute. You can talk about your day, read an article, or read the script below to help you reach the ideal length:

    "I am recording this audio to create a custom voice clone for my AI assistant. The goal is to capture the natural tone, pacing, and emotion of my voice so the AI sounds exactly like me.

    It is important that I speak clearly and maintain a consistent volume throughout this recording. I shouldn't speak too fast, nor too slow, but rather at a comfortable, conversational pace. If I want my AI to sound energetic, I need to sound energetic right now. If I want it to sound calm and professional, I should speak calmly and professionally.

    The technology behind this works by analyzing the unique characteristics of my speech—the inflections, the pitch, and the slight pauses between sentences. By giving it around a minute or two of high-quality audio without any background noise, echo, or interruptions, the system will have everything it needs to generate a highly realistic synthetic voice."