Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/bharathajjarapu/voicecipher
Local Speech transcription
https://github.com/bharathajjarapu/voicecipher
transformerjs whisper
Last synced: 2 months ago
JSON representation
Local Speech transcription
- Host: GitHub
- URL: https://github.com/bharathajjarapu/voicecipher
- Owner: bharathajjarapu
- License: mit
- Created: 2024-07-31T14:49:38.000Z (5 months ago)
- Default Branch: main
- Last Pushed: 2024-08-17T17:47:13.000Z (4 months ago)
- Last Synced: 2024-09-26T06:21:45.751Z (3 months ago)
- Topics: transformerjs, whisper
- Language: JavaScript
- Homepage: https://TinyUrl.com/VoiceCipher
- Size: 80.1 KB
- Stars: 2
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# VoiceCipher 🎙️
Local Speech transcription using Tiny Whisper modelWelcome to **VoiceCipher**, a static web application that uses [transformer.js](https://cdn.jsdelivr.net/npm/@xenova/transformers) to run the Tiny Whisper model directly through JS on the user's local machine to transcribe audio to English only.
## Features 🌟
- **User-Friendly Interface**: Clean and intuitive design for a seamless experience. 🎨
- **Flexible Input Options**: Transcribe from URLs, file uploads, or direct recordings. 🔗📁🎤
- **Speech-to-Text Transcription**: Convert your voice recordings into text using state-of-the-art ML models. 🧠## Getting Started 🚀
To get started with VoiceCipher, follow these steps:
1. **Clone the Repository**
```bash
git clone https://github.com/bharathajjarapu/voicecipher.git
```2. **Navigate to the Project Directory**
```bash
cd voicecipher
```3. **Go Live or Run the HTML File**
Simply open `index.html` in your browser to launch the application. 🌐## Usage 🛠️
1. **Choose Your Input Method**:
- **From URL**: Enter the URL of the audio file. 🌐
- **From File**: Upload an audio file from your computer. 📁
- **Record**: Record audio directly through your microphone. 🎤2. **Transcribe**: Click the "Transcribe" button to start converting your audio into text. ✏️
3. **View Results**: The transcription will appear in the text area below the buttons. Note that it is currently limited to 20 seconds for most devices; improvements are planned. ⏱️
## How It Works 🔍
VoiceCipher uses **[transformers.js](https://cdn.jsdelivr.net/npm/@xenova/transformers)** to run the Whisper model for speech-to-text conversion. Here's a brief explanation:
- **Transformer.js**: This library is used to interface with the Whisper model, enabling speech recognition directly in the browser. 📊
- **Static Execution**: Unlike many ML models that rely on WebAssembly for execution, Whisper runs statically without WebAssembly. This approach simplifies deployment and ensures compatibility across different environments. ⚙️## Dependencies 📦
- [Bootstrap](https://getbootstrap.com/) for styling. 🎨
- [Xenova Transformers](https://cdn.jsdelivr.net/npm/@xenova/transformers) for speech recognition. 🧠## Future Ideas 🚧
1. Adding support for audio clips longer than 20 seconds for all devices.
2. Adding a multilingual model to transcribe various languages, like Telugu. 🌍## License 📝
This project is licensed under the MIT License. See the [LICENSE](LICENSE) file for details.
## Contact 📫
For any questions or contributions, please contact [[email protected]](mailto:[email protected]).
Made using [Transformer.js](https://github.com/xenova/transformers.js) 🚀