bol-ke-dikha

A zero-dependency, formant-based speech synthesiser for Node.js. Converts text to audio using built-in OS playback tools — no external libraries required.

Just run

npx bol-ke-dikha

OR

npx bol-ke-dikha my name is hello

Requirements

• Node.js (CommonJS)
• OS audio player: aplay (Linux), afplay (macOS), or PowerShell SoundPlayer (Windows)

Usage

CLI

# Demo
node bol-ke-dikha.js

# Speak a word
node bol-ke-dikha.js hello

# Speak a sentence
node bol-ke-dikha.js hello world

# Interactive REPL
node bol-ke-dikha.js --repl

REPL Commands

| Input | Action | |---|---| | Any text | Speaks the words | | --letters ABC | Spells out each letter by name | | exit / quit / Ctrl+C | Exits |

Module

const { speakWord, playAlphabets, startSpeakPrompt } = require('./bol-ke-dikha');

speakWord('hello world');       // speaks words
speakWord('42');                // speaks "four two"
playAlphabets('ABC');           // spells "ay bee see"
startSpeakPrompt();             // starts interactive REPL

How It Works

Synthesises speech by generating harmonic audio at a 130 Hz fundamental pitch, shaped by formant filters (F1/F2/F3) to approximate vowel sounds. Consonants use noise bursts or approximants. Audio is written as a temporary WAV file and played via the OS, then deleted.

Supported Vocabulary

• Common English words, numbers (0–9, multi-digit), letters (A–Z)
• Tech terms: chat, gpt, claude, deep, seek, giga, mega, hertz, chitti
• Unknown words fall back to letter-by-letter phoneme sounds

License

ISC