Company Overview & Pitch Deck

Deploy AI Agents
in Minutes, Not Months

myParrot.live is a multi-tenant AI agent platform that lets any business build, brand, and deploy production-ready chat, voice, SMS, and avatar agents — with zero ML expertise.

Start Free TodayRead User Guide ↓
The Problem

Businesses are drowning in repetitive conversations

Customer support, lead qualification, onboarding, bookings — 70% of business conversations follow predictable patterns. Yet most companies still handle them manually, at enormous cost.

73%of customer inquiries are the same questions asked repeatedly
$8B+spent annually on human agents handling automatable tasks
8 minaverage wait time customers tolerate before abandoning
3-6 motypical enterprise AI deployment timeline with custom builds
The Solution

An AI workforce that works 24/7, scales instantly, and never takes sick days

myParrot provides a fully managed platform to build, customize, and deploy AI agents across every customer-facing channel — from your website chat widget to phone voice calls to avatar video kiosks.

  • No AI/ML expertise required — template-driven setup
  • Live in production in under 10 minutes
  • Multi-channel: text, voice, SMS, avatar video, Telegram
  • RAG knowledge base — trained on your documents, not the internet
  • White-label ready — your brand, your domain
  • Enterprise-grade multi-tenancy with per-tenant usage controls
💬

Text Chat

Deploy a smart chat agent on your website in minutes

🎙️

Voice Calls

Natural phone conversations with Azure neural voices

📱

SMS Agent

Two-way Twilio SMS conversations at scale

🧑‍💻

Avatar AI

Real-time video avatar for kiosks and virtual reps

🧠

RAG Knowledge

Train agents on your PDFs, docs, and websites

📊

Analytics

Full conversation logs, usage meters, and billing

How It Works

From zero to deployed in three steps

No APIs to wire up. No infrastructure to manage. Just configure and launch.

1

Choose a Template

Pick from 15+ industry templates — Healthcare, Legal, Retail, Airline, Finance — or start from scratch with a blank agent. Set the persona, tone, and knowledge in plain English.

2

Train & Configure

Upload PDFs, paste URLs, or connect your documents to the RAG knowledge base. Choose voice, avatar, or text mode. Set language, session limits, and escalation rules.

3

Embed & Launch

Copy a single <script> tag to add a chat widget to any website. Share the public URL. Connect Twilio for SMS or Azure for phone. Your agent is live.

Product

Four agent types. One platform.

Choose the right modality for every use case and audience.

💬

Text Agent

Instant answers, zero wait

Website chat widget, inline embed, or public shareable URL. Supports RAG retrieval, session memory, file attachments, and multi-language.

Common use cases

Customer supportLead qualificationFAQ botsProduct finderOnboarding flows
🎙️

Voice Agent

Human-quality phone conversations

Azure neural voices with real-time speech recognition. Handles inbound and outbound calls. Transcripts logged automatically.

Common use cases

Appointment bookingIVR replacementCold outreach follow-upDelivery updatesCollections
📱

SMS Agent

Meet customers where they text

Two-way Twilio SMS conversations. Supports long-form messages, keyword triggers, opt-out management, and campaign blasting.

Common use cases

Appointment remindersDelivery notificationsLead nurtureEmergency alertsLoyalty rewards
🧑‍💻

Avatar Agent

The face of your brand, always on

Real-time video avatar synthesis via Azure. Train on your own face photo or use preset characters. Perfect for kiosks, virtual showrooms, and VIP concierge.

Common use cases

Airport/hotel kiosksVirtual showroomAirline conciergeBank onboardingRetail assistant
Deployment

Deploy anywhere customers are

One agent, infinite channels. Every deployment is live in minutes.

🌐

Chat Widget

Floating chat bubble on any website with a single <script> tag.

📄

Inline Embed

Embed the agent directly inside a page section or div element.

🔗

Public Page

shareable link — myparrot.live/a/your-agent-slug

📱

SMS / Twilio

Two-way text messaging on any Twilio-connected number.

📞

Voice Call

Azure-powered real-time voice over WebRTC or phone bridge.

✈️

Telegram Bot

Native Telegram bot integration via BotFather webhook.

🖥️

Kiosk / Avatar

Full-screen avatar video mode for digital signage.

🔌

REST API

Call any agent programmatically via authenticated API endpoint.

📧

Email (coming)

Inbound email parsing with agent-drafted reply flows.

Pricing

Transparent pricing that scales with you

Start free. No credit card required. Upgrade when you grow.

Free

$0forever

Try the platform. No commitment.

  • 1 agent
  • 100 messages/month
  • Text chat only
  • Public page embed
  • Widget embed
Get Started

Starter

$239/month

$191/mo billed annually

For small teams automating customer conversations.

  • 5 agents
  • 2,000 messages/month
  • 100 voice minutes/month
  • RAG knowledge base
  • All embed channels
Start Starter

Pro

$399/month

$319/mo billed annually

Full AI workforce including avatar agents.

  • 20 agents
  • 10,000 messages/month
  • 500 voice minutes/month
  • 150 avatar minutes/month
  • RAG + Avatar AI
  • Priority support
Start Pro

Enterprise

Custom

Unlimited scale, custom integrations, and dedicated SLA.

  • Unlimited agents
  • Unlimited messages
  • Custom voice & avatar limits
  • Dedicated support
  • Custom pricing
  • SLA agreement
Contact Sales
Overage rates (paid plans only):
$0.02 / message — Extra messages
$0.25 / min — Extra voice minutes
$0.99 / min — Extra avatar minutes
Market Opportunity

A $47B market underserved by complexity

The global conversational AI market is projected to reach $47B by 2029. Today, 90% of businesses want AI agents but can't deploy them — because every alternative requires months of custom development.

$47B

Conversational AI market by 2029

CAGR: 23.5%

87%

of SMBs cannot afford bespoke AI development

Source: Gartner 2024

4.2×

ROI reported by companies using AI for customer service

McKinsey 2024

Customer ROI Calculator

A typical Pro-plan customer with one voice agent handling 500 calls/month:

Human agent cost (500 calls × $8/call)$4,000/month
myParrot Pro plan$399/month
Net monthly saving$3,601/month
Annual saving$43,212/year

Who We Built This For

  • 🏥

    Healthcare

    Patient intake, appointment reminders, care navigation

  • ✈️

    Airlines & Travel

    Flight booking, disruption management, loyalty support

  • 🏦

    Financial Services

    KYC onboarding, loan inquiries, account support

  • 🛒

    Retail & E-commerce

    Order support, returns, product recommendations

  • ⚖️

    Legal Services

    Intake triage, matter routing, appointment scheduling

  • 🏨

    Hospitality

    Booking changes, in-stay requests, guest feedback

Technology

Enterprise-grade infrastructure, zero vendor lock-in

Built on battle-tested Azure and OpenAI services with a cloud-agnostic architecture.

AI Models

  • · Azure OpenAI (GPT-4o mini)
  • · GPT-4 (Pro/Enterprise)
  • · Custom LLM policies per tenant

Voice & Avatar

  • · Azure Neural TTS
  • · Azure Speech STT
  • · Azure Avatar real-time synthesis
  • · WebRTC streaming

Backend

  • · NestJS (Node.js)
  • · PostgreSQL + Prisma
  • · RAG with pgvector
  • · Stripe billing

Frontend

  • · Next.js 14 App Router
  • · Tailwind CSS
  • · Widget (Shadow DOM)
  • · Azure CDN (edge)

Documentation

User Guide

Step-by-step instructions for everything you can do on myParrot.live — from your first agent to enterprise deployments.

Step 1

Getting Started

Create your account

  1. Go to myparrot.live/register
  2. Enter your full name, work email, and password
  3. Verify your email address (check spam if needed)
  4. You're automatically on the Free plan — no credit card required
💡
Free plan includes: 1 agent, 100 messages/month, text chat only, widget and public page embed. Upgrade any time from the Billing page.

Dashboard overview

After login you land on the Overview dashboard. The sidebar gives you access to:

  • 📊 OverviewUsage summary and recent activity
  • 🤖 All AgentsManage and deploy your AI agents
  • ➕ New AgentCreate a new agent from scratch or template
  • 📚 Knowledge BaseUpload documents for RAG retrieval (Starter+)
  • 💬 ConversationsAll chat logs across every agent
  • 📈 UsageDetailed usage metrics and billing period stats
  • 🎙️ Voice StudioCustom voice and avatar training
  • 💳 BillingPlan management, invoices, and upgrades
Step 2

Creating Your First Agent

From the sidebar, click "New Agent"

  1. Name your agent — the display name visible to users
  2. Choose a type: Text, Voice, SMS, or Avatar
  3. Pick a template (optional) — 15+ industry templates pre-fill the system prompt and configuration for your use case
  4. Write the system prompt — define the agent's role, personality, scope, and escalation rules in plain English
  5. Configure voice/avatar (Voice/Avatar types) — choose a neural voice from the dropdown
  6. Attach knowledge sources (Starter+) — link documents from your Knowledge Base for RAG answers
  7. Click Create Agent
🤖
What goes in the system prompt? Write it as if you're briefing a new employee. Example:

"You are Jordan, a customer support specialist for AcmeCo. Help customers with order tracking, returns, and product questions. Always be friendly and concise. If the customer is angry, escalate to a human by collecting their email and issue summary."

Available Templates

HealthcareReal EstateFinancial ServicesRetailLogisticsEducationHospitalityLegalInsuranceAirline Flight BookingAirline DisruptionCheck-In & BoardingLoyalty & MilesManufacturing
⚠️
Free plan limit: 1 agent maximum. Upgrade to Starter (5 agents) or Pro (20 agents) to add more. You'll see a clear error message when the limit is reached.
Step 3 · Starter+

Training with the Knowledge Base

Upload your documents and your agent answers questions from your actual content — not the public internet.

Supported file types

  • PDF documents (product manuals, policies, FAQs)
  • Plain text (.txt) and Markdown files
  • CSV data files (product catalogs, pricing sheets)
  • DOCX Word documents
  • Web URLs (paste a URL to scrape the page content)

How to add knowledge

  1. Go to Knowledge Base in the sidebar
  2. Click Add Source
  3. Upload a file or paste a URL
  4. Wait for status to show READY (usually under 30 seconds)
  5. Open the agent you want to train → Edit → Knowledge Sources → link the source

How RAG works

1

Your document is split into chunks and converted to vector embeddings using Azure OpenAI.

2

When a user asks a question, the agent retrieves the most relevant chunks from your knowledge base.

3

The top chunks are injected into the LLM context alongside the user question.

4

The agent generates a precise, grounded answer — not a hallucination.

📌
Plan limits: Free plan = no RAG. Starter+ = 20 RAG chunks per query. For best results, keep each source focused on a single topic and keep chunks under 500 words.
Step 4

Testing & Deploying Your Agent

Test before you go live

From the Agents list, click Test Chat on any agent. This opens an in-platform console where you can chat directly with the agent under your identity. All test conversations are logged under the TEXT channel.

Publish the agent

An agent is private until you publish it. Once published, it gets a unique public slug (e.g. my-support-bot). You can then:

  • Share the public URL: myparrot.live/a/{slug}
  • Copy the widget snippet and paste it into your website's HTML
  • Use the inline embed code for a div-based embed
  • Generate and rotate the API key for programmatic access

Widget installation (30 seconds)

Paste this before </body> in your HTML:

<script
  src="https://myparrot.live/widget.js"
  data-agent="your-agent-slug"
  defer
></script>

Inline embed:

<div
  data-myparrot-agent="your-agent-slug"
  style="height:500px"
></div>
<script src="https://myparrot.live/widget.js" defer></script>
🔒
The widget uses Shadow DOM isolation — it won't conflict with your existing CSS. No API key is needed for the widget; the public slug acts as the identifier.
Step 5

Conversation Analytics

Every message, from every channel, is captured and searchable.

Where to find conversations

  • Conversations (sidebar) — all conversations across all agents, filterable
  • Agents → Chats button — conversations for a single agent only

Channel types

Test ChatFrom the in-platform test console
Widget / PublicFrom the embed widget or public URL
VoiceVoice call conversations
Public VoiceVoice from the public page
SMSTwilio SMS two-way conversations

Reading a conversation

Click any row in the conversation list to open the full message thread on the right. Each conversation shows:

  • Channel badge (where it came from)
  • Start time and message count
  • Visitor reference (IP slug, phone number, or identifier)
  • Full message thread with timestamps
  • Voice input indicator (🎙) for audio-transcribed messages
🔍
Use the search bar to filter by channel, session ID, or visitor reference. Pagination loads 20 conversations per page — use Prev/Next to browse older sessions.
Advanced · Starter+

Voice Studio

Create custom neural voices and photo-realistic avatar characters trained on your own recordings.

Custom Voices

  1. Go to Voice Studio in the sidebar
  2. Click New Voice
  3. Record or upload 50–200 audio samples (5–15 seconds each)
  4. Provide text transcripts for each sample
  5. Submit for Azure Custom Neural Voice training (24–48 hours)
  6. Once READY, assign the voice to any agent
📋
Consent requirement: Azure requires a spoken consent recording from the voice talent stating they agree to the synthetic voice model. The platform will prompt you for this during setup.

Custom Avatar (Photo)

  1. Click New Avatar → choose PHOTO type
  2. Upload a clear, front-facing photo (min 300×300px, neutral background)
  3. Give consent and submit
  4. Azure processes the photo and applies it to a prebuilt character
  5. Assign the avatar to any Avatar-type agent

Custom Avatar (Full Video)

Record a full-body training video (2–5 minutes, controlled lighting, neutral background). Submit for Azure training — takes 2–3 business days. Full-body avatars support gestures and expressions. Requires Pro or Enterprise plan.

Billing

Managing Your Subscription

Upgrading your plan

  1. Go to Billing in the sidebar
  2. Choose Monthly or Annual (save 20%)
  3. Click Upgrade to Starter or Upgrade to Pro
  4. Complete checkout via Stripe (credit card or bank debit)
  5. Your plan upgrades instantly after payment

Managing your subscription

Click Manage Subscription to open the Stripe Customer Portal where you can:

  • Switch between monthly and annual billing
  • Upgrade or downgrade your plan
  • Update your payment method (card or bank)
  • Cancel your subscription (access continues until period end)
  • Download past invoices as PDF

Usage meters

The Billing page shows four usage meters for the current billing period:

Messages

Total messages across all agents and all channels. Resets on your billing renewal date.

Agents

Total active agents in your workspace. This is a hard limit, not a monthly counter.

Voice Minutes

Total minutes of voice call audio processed this period.

Avatar Minutes

Total minutes of live avatar video synthesis this period.

📊
Overages (paid plans only): If you exceed included limits on a paid plan, usage continues at overage rates — $0.02/msg · $0.25/voice min · $0.99/avatar min. Free plan blocks at 100 messages/month.
FAQ

Frequently Asked Questions

Can I change my agent type after creating it?

Yes. Go to Agents → Edit your agent → change the type. Note that if you switch to AVATAR, you need a Pro plan or higher.

Do conversations expire or get deleted?

No. All conversation history is retained indefinitely. Only your plan's active usage is tracked for billing purposes.

Can I white-label the widget with my own branding?

Yes. The widget shows your agent's name and can be styled via CSS variables. Contact sales for full white-label options including custom domain.

What happens when I hit the free tier 100 message limit?

The agent will return a friendly message indicating the limit has been reached and prompting the user to contact you. Upgrade to any paid plan to restore access immediately.

Can multiple team members access the same workspace?

Yes. Go to Admin (sidebar) to invite team members. Each user gets a role: CLIENT (admin access) or CUSTOMER (limited access). User counts don't affect billing.

What AI model does myParrot use?

By default, GPT-4o mini (Azure OpenAI) for speed and cost efficiency. SUPER_ADMINs can configure GPT-4 or other approved models per tenant via the platform admin panel.

How do I set up a Telegram bot?

Go to Integrations → Telegram → follow the BotFather steps to get a bot token, then paste it in the platform. The agent will immediately start responding to Telegram messages.

Is my data private and secure?

Yes. Each tenant's data is fully isolated at the database level with tenant-scoped queries. Knowledge base documents and conversations are stored in your dedicated Azure PostgreSQL instance. No data is shared between tenants.

Can I use myParrot for outbound calls or SMS?

SMS: Yes, you can trigger outbound SMS via the REST API. Outbound voice calls: available on Enterprise plan. Contact sales for campaign setup.

What is the Enterprise plan?

Enterprise is custom-priced based on your specific usage requirements — agents, message volume, voice/avatar minutes, and dedicated support SLA. Pricing is set by our team. Click "Contact Sales" to start the conversation.

Ready to deploy your first agent?

Start free, upgrade when you grow. No credit card required on the free tier.

Create Free Account →Contact Sales