Disclaimer

Last updated: March 10, 2026

AI Transcription Accuracy

Whisper STT uses OpenAI's Whisper AI model for speech recognition and translation. While Whisper is one of the most advanced speech recognition systems available, no AI transcription system is 100% accurate.

Factors that can affect accuracy include:

  • Audio quality (background noise, microphone quality, recording conditions)
  • Speaker accent, dialect, or speech patterns
  • Multiple overlapping speakers
  • Technical terminology, proper nouns, and specialized vocabulary
  • Low-resource languages with less training data
  • Music, sound effects, or other non-speech audio

Not a Substitute for Professional Services

Whisper STT's transcriptions should not be used as a substitute for professional transcription services in situations where accuracy is critical, including:

  • Legal proceedings: Court transcripts, depositions, and legal documents
  • Medical records: Clinical notes, diagnoses, and medical documentation
  • Financial documents: Financial reports, audit transcripts, and regulatory filings
  • Academic submissions: Research papers, dissertations, and official academic records

Always review and verify AI-generated transcriptions before using them for important purposes.

Translation Limitations

The translation feature converts speech in any supported language to English text. Machine translation may not capture nuances, idioms, cultural context, or specialized terminology accurately. For important communications, consult a professional translator.

Device and Browser Requirements

Whisper STT runs AI models directly in your browser, which requires:

  • A modern web browser (Chrome, Firefox, Edge, or Safari)
  • Sufficient RAM (at least 4 GB recommended)
  • Storage space for the model cache (~150 MB)
  • Internet connection for the initial model download

Performance varies by device. Older or less powerful devices may experience slower processing times.

No Warranty

The Service is provided without any warranty, express or implied. We do not guarantee the accuracy, reliability, or completeness of any transcription or translation produced by the Service.

Third-Party Content

The Whisper model is developed by OpenAI. Transformers.js is developed by Hugging Face. Whisper STT is not affiliated with, endorsed by, or sponsored by OpenAI or Hugging Face. We are not responsible for the functionality, performance, or limitations of these third-party components.

Contact

For questions about this disclaimer, contact us at legal@whisperstt.com.