← Back to ToolsIQ
Beta
Voice AI Tool

EchoSpellIQ

Your AI voice should never butcher a name.

Say a name. Hear the AI say it back. Iterate until perfect. One phone call sets the pronunciation for every future interaction.

The Problem

Names Are Personal. Mispronunciation Is Unprofessional.

Every voice AI system butchers names. Patients hang up. Clients lose trust. Staff cringe. The pronunciation problem is universal — and until now, unsolved.

72%
Feel Disrespected

of people say mispronouncing their name feels dismissive or disrespectful, especially from a service they're paying for.

0
TTS Engines Get It Right

No text-to-speech engine correctly pronounces Nguyen, Bhatia, Wojciechowski, or Siobhan out of the box. Not one.

1st
Impression Matters

When your AI receptionist says "Hello, Dr. Buh-HAT-ee-uh" instead of "Dr. BAH-tya", the patient's trust drops before the conversation starts.

390+
Pre-built Pronunciations
Names from 15+ language families
<30s
Per Name
Average time to approve a pronunciation
0ms
Downtime
Hot-reload — no server restarts needed
100%
Fidelity
Test with the same engine callers hear
How It Works

Six Steps to Perfect Pronunciation

Call the number, say the name, hear it back. No software to install, no phonetic alphabet to learn. Just your voice and your phone.

Step 1
Call the Number

Call your dedicated EchoSpellIQ number and enter your practice PIN. Or use the web interface for batch setup.

Step 2
Say the Name

Speak the name naturally — "Dr. Bhatia", "Xiomara", "Wojciechowski". Our STT captures exactly what you said.

Step 3
AI Generates Phonetic

Gemini AI produces a TTS-optimized phonetic spelling — simple English syllables with stress markers, not IPA gibberish.

Step 4
Hear It Back

The system reads the phonetic spelling back to you using the SAME voice engine your callers will hear. No surprises.

Step 5
Iterate or Approve

Not right? Give a hint — "stress the first syllable" or "it rhymes with car" — and we regenerate. Loop until perfect.

Step 6
Live Instantly

Approved pronunciations are pushed to your voice server immediately. No restarts, no deployments. The next call gets it right.

Use Cases

Built for Voice AI Teams

Whether you're onboarding a dental practice, deploying a call center, or shipping a voice assistant — pronunciation accuracy is table stakes.

Practice Onboarding

Before going live, run every doctor, hygienist, and staff name through the loop. Five minutes per name, done once, correct forever.

Live Self-Correction

When a caller says "that's not how you say my name," the system captures the correction in real-time and updates instantly.

Multi-Language Names

Chinese, Vietnamese, Arabic, Polish, Gaelic — our AI handles names from every language family, not just English-friendly ones.

Any Voice AI System

Export pronunciations as a simple text file compatible with any TTS engine — NVIDIA Riva, Edge TTS, ElevenLabs, or your custom stack.

Included

Everything in the Box

Dedicated phone number for pronunciation setup
390+ pre-built pronunciations (15+ language families)
AI-powered phonetic generation via Gemini
Real-time TTS preview over the phone
Iterate with voice feedback until perfect
Hot-reload to your voice server (no restarts)
Web dashboard for batch management
Export as TXT, JSON, or SSML format
Live-call self-correction (caller says "that's wrong")
API access for custom integrations
Works with any TTS engine (Riva, Edge, ElevenLabs)
Unlimited names per month
Pricing

Simple, Flat Pricing

One price. Everything included. No per-name charges.

Everything Included
$10/month

per practice or project

Unlimited pronunciation entries
Dedicated setup phone number
AI phonetic generation
Real-time TTS preview
Voice server hot-reload
Web dashboard access
API & export tools
Live-call self-correction
Start 14-Day Free Trial

No credit card required. Cancel anytime.

VoicesIQ customers get EchoSpellIQ included with their plan at no extra cost.

Stop Mispronouncing Names.

Your AI should say every name as well as a human would. EchoSpellIQ makes that happen in under a minute per name.