Built on Sarvam AI
A phone receptionist built on Sarvam AI.
Sarvam AI makes speech models for Indian languages. Lexora Voice puts them on a real phone line: a receptionist that hears callers with Sarvam's Saaras speech recognition, speaks with Sarvam's Bulbul voice, and books appointments for your business.
Why Sarvam AI for a phone line
Phone calls in India are code-mixed: Bengali or Hindi with English words through them. Saaras is built for Indian speech, English words and all, and detects the language the caller is speaking. Bulbul speaks Bengali and Hindi with native voices instead of an English voice guessing at the words.
What we built on top
A model is not a receptionist. Lexora Voice adds the rest: the phone line, your business's facts, booking against your hours, reading numbers and times out the way people say them, following the caller between languages, and an email with the transcript after every call.
Ready to use, no code
You don't need a Sarvam account, an API key or a developer. Forward your business number to us, or take a new one, tell us your fees and timings, and it answers.
Hear it first
The demo line runs the real product. Call it and speak Bengali, Hindi or English.
Questions
Is Lexora Voice made by Sarvam AI?
No. Lexora Voice is made by Lexora Technologies in Kolkata. We use Sarvam AI's speech models; we are not affiliated with Sarvam.
Which Sarvam AI models does it use?
Saaras for speech recognition and Bulbul for the voice.
Do I need my own Sarvam AI API key?
No. Everything runs on our side. You only need a phone number.
Don't take our word for it. Call it.
This is a live line. Ask it what your customers ask you.