Snomi.ai
How I've build voice cloning & text-to-speach app from scratch in 2 days to read bedtime stories to my little son.
Prototyping
Vibe Coding
UI Design
Overview
Snomi.ai is a personal project that grew naturally out of my own life. When I first became a dad, I often struggled to stay awake while reading bedtime stories after long workdays. I wanted my son to still hear my voice on the nights I couldn’t keep my eyes open.
Role
All–in–one
Years
2024
Type
Side Project
Team
My son Franek
Tools
Cursor, ElevenLabs, TTS Models, Figma
The Challange
Overview
I realised this is a common problem for parents — trying to stay present for their kids at the end of a long workday, especially during bedtime stories. I saw an opportunity to solve that.
Goal #1
A fully working prototype of a platform where parents can clone their voice and have bedtime stories read to their kids.
Goal #2
Learn vibe coding tools on the way.
Discovery
Audio LLM's Research
As the first step I've researched available audio LLM's that allow to voice cloning. I've tested popular platforms like ElevenLabs, Minimax, Resemble.ai.
Vibe Coding Tools Research
After hearing all the hype around tools like Cursor, Replit, Lovable, and v0, I wanted to try them myself. They blew me away — I managed to prototype the whole flow in minutes.
What I've acomplished:
Researched & choose ElevensLab's audio LLM
Vibe coded full voice cloning flow using Cursor
Successfully connected ElevenLabs API
Read few stories to my Son through Snomi.ai
Design Excerise
Having a working prototype pushed me to design a clean, simple UI for the product. I built it exactly how I imagined it — straightforward, stylish, and easy to use.
