> ## Documentation Index
> Fetch the complete documentation index at: https://api-docs.ollang.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Whisper

**Whisper** is an automatic speech recognition (ASR) system developed by OpenAI. Whisper is trained on a large and diverse dataset of multilingual and multitask supervised data collected from the web, making it robust and versatile for various speech recognition tasks.

## Key Capabilities

* **Multilingual Support**: Whisper supports numerous languages, allowing it to transcribe speech from diverse linguistic backgrounds.
* **Robust Performance**: It is capable of handling different acoustic settings, including noisy environments and varied accents.
* **Automatic Language Detection**: The model can automatically detect the language spoken in the audio input.
* **Versatility**: Suitable for transcribing lectures, meetings, podcasts, conversations, and more.
* **Open Source**: Available on GitHub, enabling developers to access, modify, and contribute to the codebase.

## Metrics

* **WER (Word Error Rate)**: Whisper demonstrates a low word error rate across multiple languages and benchmarks, indicating high transcription accuracy.
* **Languages Supported**: Over 50 languages.
* **Training Dataset**: 680,000 hours of multilingual and multitask supervised data.

## Use Cases

1. **Transcription Services**: Automating the conversion of audio files into text for uses such as subtitles, meeting notes, and academic research.
2. **Language Translation**: In combination with translation models, Whisper can facilitate real-time speech translation.
3. **Accessibility Tools**: Enhancing accessibility for individuals with hearing impairments by providing real-time captions for spoken content.
4. **Voice-Activated Assistants**: Serving as the core technology for more responsive and accurate voice-activated user interfaces.