You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

Hinglish Code-Switched Conversational Dataset v1

Overview

This dataset contains structured Hinglish conversational voice data built to reflect how people actually speak in real-world interactions.

Most speech datasets are clean, scripted, or heavily processed. That works in controlled testing, but it breaks in production where speakers interrupt each other, switch languages, use regional accents, pause mid-thought, and shift context naturally.

This sample release demonstrates Sonexis’ structured approach to conversational voice data.

What This Dataset Contains

Each conversation includes:

  • speaker-separated audio
  • structured transcripts
  • conversation-level metadata
  • speaker-level metadata
  • code-switching indicators
  • conversational annotations
  • scenario-level context

The dataset is designed for teams building or evaluating multilingual voice AI systems for India and other multilingual markets.

Languages

This release focuses on Hinglish, with natural switching between:

  • Hindi
  • English

The conversations may include informal speech, regional accent variation, interruptions, overlaps, and context shifts.

Dataset Structure

audio/
  conversation_0001/
    conversation_0001_spk1_16k.wav
    conversation_0001_spk2_16k.wav
  conversation_0002/
    conversation_0002_spk1_16k.wav
    conversation_0002_spk2_16k.wav

transcripts/
  conversation_0001.json
  conversation_0002.json

annotations/
  conversation_0001.json
  conversation_0002.json

manifests/
  conversations.jsonl
  speakers.jsonl
  speaker_separated.jsonl
  utterances.jsonl
  final_transcripts.jsonl

metadata.jsonl
conversation_metadata.json

docs/
  dataset_schema.json
  schema.md
Downloads last month
35