ConvoAI Builder

AI-powered ConvoAI project builder with multi-persona assistance

ConvoAI Builder is an intelligent CLI tool that helps you build, configure, and deploy ConvoAI projects through natural conversation with specialized AI personas.

Features

🤖 Multi-Persona AI System - Get expert help from 5 specialized AI personas:

• 🏗️ Architect - System design & architecture decisions
• 💻 Developer - Code implementation & debugging
• ⚙️ Configurator - Setup & configuration
• 🚀 Deployer - Deployment & DevOps
• 🎓 Advisor - Best practices & guidance

✨ Interactive Project Setup - Initialize projects with guided setup

🎯 Context-Aware Assistance - Maintains project state across conversations

📦 Template System - Pre-built agent templates for common use cases

🚀 Multi-Platform Deployment - Deploy to Vercel, AWS, GCP, Docker, or self-hosted

Installation

npm install -g convoaibuilder

Quick Start

1. Initialize a New Project

convoaibuilder init my-convoai-app
cd my-convoai-app

2. Configure Your Project

# Copy environment template
cp .env.example .env

# Edit .env and add your API keys
# - OPENAI_API_KEY
# - ELEVENLABS_API_KEY
# - AGORA_APP_ID
# - AGORA_APP_CERT

3. Chat with the Builder Agent

convoaibuilder chat

Example conversation:

You: I want to create a customer support agent

💻 developer: I'll help you create a customer support agent.
Let me generate the code...

[Generates agent code with configuration]

✓ Agent created: backend/src/agents/customer-support.ts
✓ Updated: convoai.config.yaml

4. Add More Agents

convoaibuilder add-agent

5. Deploy Your Project

convoaibuilder deploy

Commands

convoaibuilder init [project-name]

Initialize a new ConvoAI project with interactive setup.

Options:

• -t, --template <type> - Project template (express, nextjs, fullstack)
• -y, --yes - Skip prompts and use defaults

Example:

convoaibuilder init my-app --template fullstack

convoaibuilder chat

Start interactive chat with ConvoAI Builder Agent.

Options:

• -p, --persona <name> - Start with specific persona
• -w, --workflow <name> - Load specific workflow

Example:

convoaibuilder chat --persona architect

Special Commands in Chat:

• /help - Show available commands
• /personas - List available personas
• /workflows - List available workflows
• /context - Show current context
• /reset - Reset conversation context
• exit - Exit chat mode

convoaibuilder add-agent

Add a new conversational agent to your project.

Options:

• -n, --name <name> - Agent name
• -t, --template <type> - Agent template

Example:

convoaibuilder add-agent --name "Sales Assistant" --template sales

convoaibuilder deploy

Deploy your ConvoAI project.

Options:

• -p, --platform <name> - Deployment platform

Supported Platforms:

• docker - Local Docker containers
• vercel - Vercel (recommended for quick start)
• aws - Amazon Web Services
• gcp - Google Cloud Platform
• self-hosted - Generate deployment files

Example:

convoaibuilder deploy --platform vercel

convoaibuilder info

Show ConvoAI Builder Agent information and available personas.

AI Personas

ConvoAI Builder uses specialized AI personas that adapt to your needs:

🏗️ Architect

Expert in system design and architecture decisions. Asks clarifying questions about your requirements and recommends optimal technology stacks and provider configurations.

Use for: Architecture decisions, scalability planning, provider selection

💻 Developer

Full-stack developer specialized in ConvoAI implementations. Generates clean, type-safe code following best practices.

Use for: Code implementation, debugging, custom features

⚙️ Configurator

Configuration expert for ConvoAI systems. Helps set up environment variables, agent configurations, and provider settings.

Use for: Setup, configuration, environment management

🚀 Deployer

DevOps expert for deploying ConvoAI projects. Guides you through deployment to various platforms with best practices.

Use for: Deployment, CI/CD, infrastructure setup

🎓 Advisor

Mentor and guide for ConvoAI developers. Teaches concepts, best practices, and helps with learning.

Use for: Learning, best practices, troubleshooting, guidance

Agent Templates

ConvoAI Builder includes pre-built templates for common use cases:

• customer-service - Customer support and service agents
• sales - Sales and product assistance agents
• education - Educational tutoring agents
• healthcare - Healthcare assistance (HIPAA considerations)
• hr-support - HR and employee support agents
• interview-practice - Interview simulation and practice
• custom - Start from scratch

Project Structure

After initialization, your project will have this structure:

my-convoai-app/
├── backend/
│   ├── src/
│   │   └── agents/
│   └── tests/
├── frontend/
│   └── src/
├── convoai.config.yaml
├── .env.example
├── CONVOAI_LIBRARY_CONTEXT.md
├── package.json
└── README.md

Configuration

convoai.config.yaml

Main configuration file for your ConvoAI project:

version: 1.0

providers:
  openai:
    apiKey: ${OPENAI_API_KEY}
    model: gpt-4-turbo

  elevenlabs:
    apiKey: ${ELEVENLABS_API_KEY}

  agora:
    appId: ${AGORA_APP_ID}
    appCertificate: ${AGORA_APP_CERT}

agents:
  customer-support:
    name: Customer Support Bot
    template: customer-service
    language: en-US
    voice: female-warm
    features:
      transcription: true
      feedback: true
      analytics: true

Environment Variables

Required environment variables (.env):

# OpenAI API Key
OPENAI_API_KEY=sk-...

# ElevenLabs API Key
ELEVENLABS_API_KEY=...

# Agora Credentials
AGORA_APP_ID=...
AGORA_APP_CERT=...

# JWT Secret
JWT_SECRET=...

# Database (optional, auto-uses SQLite)
DB_HOST=localhost
DB_PORT=5432
DB_NAME=convoai
DB_USER=postgres
DB_PASSWORD=...

Requirements

• Node.js >= 18.0.0
• npm or yarn
• API keys for:
  • OpenAI (for LLM)
  • ElevenLabs or Azure (for TTS)
  • Agora (for RTC)

Examples

Create a Healthcare Voice Agent

convoaibuilder init healthcare-assistant
cd healthcare-assistant
convoaibuilder chat

You: I need a HIPAA-compliant patient intake agent

🏗️ architect: For HIPAA compliance, I recommend:
- Azure TTS (HIPAA compliant)
- OpenAI Business/Enterprise
- PostgreSQL with encryption
- Self-hosted deployment

🎓 advisor: Let me help you understand HIPAA requirements...

Add Multiple Agents

convoaibuilder add-agent --name "Support Bot" --template customer-service
convoaibuilder add-agent --name "Sales Bot" --template sales
convoaibuilder add-agent --name "Tutor Bot" --template education

Deploy to Production

convoaibuilder deploy --platform vercel

✓ Backend deployed: https://api-abc123.vercel.app
✓ Frontend deployed: https://app-abc123.vercel.app

Programmatic API

You can also use ConvoAI Builder programmatically:

const { ConvoAIBuilderAgent } = require('convoaibuilder');

const agent = new ConvoAIBuilderAgent({
  contextFile: './CONVOAI_LIBRARY_CONTEXT.md',
  workspaceRoot: process.cwd(),
  defaultPersona: 'advisor'
});

// Chat with the agent
const response = await agent.chat('How do I create a new agent?');

console.log(response.persona); // 'developer'
console.log(response.prompt);  // Full prompt with context

License

UNLICENSED - Proprietary software. All rights reserved.

Copyright (c) 2025 Sarthak Batra

Support

For issues, questions, or feature requests, please contact the maintainer.

Links

• ConvoAI Library Documentation
• Agora Documentation
• OpenAI API
• ElevenLabs API

Built with ❤️ by Sarthak Batra