Call2Fly AI Docs

Overview

Call2Fly AI is designed around a simple, four-phase workflow that takes you from concept to production.

Platform Overview

Call2Fly AI is designed around a simple, four-phase workflow that takes you from concept to production.

Dashboard Navigation

The dashboard has a left sidebar organized into four main sections:

SectionPurpose
BUILDCreate and manage AI agents and knowledge bases
DEPLOYConfigure phone numbers and batch call campaigns
MONITORTrack call/chat history and view analytics
SYSTEMManage billing and workspace settings

The Four Phases

1. Build

The Build phase is where you create and configure your AI agents:

Agents

  • Create voice agents (4 types) or chat agents (3 types)
  • Choose from 30+ pre-built templates
  • Configure agent behavior with conversation flows or prompts
  • Add custom functions for calendar booking, order management, etc.
  • Import/export agents as JSON

Knowledge Base

  • Upload documents (PDF, Word, TXT, Markdown)
  • Crawl and sync website content
  • Add text articles manually
  • Organize with names and descriptions

Learn more about building agents →

2. Test

Before deploying to production, test your agents thoroughly:

  • Web Demo: Try out your agents directly in your browser
  • Test Panel: Real-time testing within the agent editor
  • Recording Playback: Listen to recorded test conversations
  • Iterate: Refine your agent's responses and behavior

Learn more about testing →

3. Deploy

When you're ready, deploy your agents to handle real conversations:

Phone Numbers

  • Purchase numbers via Twilio or Telnyx
  • Connect existing numbers via SIP trunking
  • Assign inbound/outbound agents to numbers

Batch Calls

  • Upload recipient lists (CSV)
  • Schedule campaigns for specific times
  • Set execution windows and concurrency limits
  • Monitor campaign progress

Learn more about deployment →

4. Monitor

Track performance and continuously improve your agents:

Call History

  • View all calls with 13+ data columns
  • Filter by agent, date, status, sentiment, and more
  • Listen to recordings and view transcripts
  • Export to CSV

Chat History

  • Track chat sessions and messages
  • Analyze user sentiment
  • View conversation details

Analytics

  • Create custom dashboards with various chart types
  • Track metrics: call volume, duration, success rate, sentiment
  • Compare performance over time
  • Filter by agent or time period

Learn more about monitoring →

Core Concepts

Agent Types

Voice Agents

TypeBest For
Conversation FlowRigid, highly formatted conversations with defined paths
Single PromptSimple, freeform conversations
Multi-PromptFlexible conversations with structured flows
Custom LLMAttach your own LLM endpoint

Chat Agents

TypeBest For
Conversation FlowStructured text conversations
Single PromptSimple chatbot interactions
Multi-PromptComplex text-based flows

Workspaces

Workspaces allow you to organize your agents, phone numbers, and other resources:

  • Separate development and production environments
  • Manage different projects or clients
  • Control team access with roles (Admin, Member, Viewer)
  • Switch between workspaces easily

Functions

Agents can execute functions during conversations:

  • Calendar Integration (Cal.com) - Check availability, book appointments
  • Order Management - Check orders, process refunds
  • Custom Functions - Create your own API integrations

Webhooks

Integrate with your systems via webhooks:

  • Receive conversation events in real-time
  • Configure timeout and retry settings
  • Test webhook endpoints

Architecture

Call2Fly AI integrates with various services:

  • Telephony: Twilio, Telnyx, SIP trunking
  • AI Models: Advanced language models for natural conversation
  • Storage: Secure storage for call recordings and transcripts
  • Analytics: Real-time and historical data analysis
  • Notifications: Firebase Cloud Messaging for push notifications

Next Steps

  1. Create your first agent
  2. Add knowledge to your agent
  3. Test your agent
  4. Deploy to production
  5. Monitor performance

On this page