Welcome to the TTS Dataset Recorder! This tool helps you quickly create a high-quality dataset for Text-to-Speech (TTS) models. Whether you're starting from scratch or have a pre-existing set of text data, this app lets you record audio samples and export them with the corresponding metadata.
How to Use?
Upload a JSON File containing the sentences you'd like to record (or manually input them through the app).
Record Audio for each sentence. The app will automatically associate your recordings with the correct text.
Export the Dataset as a JSON file or Sync to HuggingFace for easy sharing and use.
Data Input Format
Your JSON file should follow this structure:
[
{ "id": "001", "text": "Hello, how are you?" },
{ "id": "002", "text": "This is a sample sentence." }
]