Whisper
OpenAI's open-source speech recognition model that transcribes audio with high accuracy across languages.
Definition
Whisper is OpenAI's open-source automatic speech recognition (ASR) model, trained on 680,000 hours of multilingual audio. It transcribes speech to text with accuracy approaching human transcriptionists, handles multiple languages, and can be run locally without API costs.
Released with open weights, Whisper has become the foundation for countless transcription applications and voice interfaces.
Why It Matters
High-quality speech recognition was historically expensive and API-dependent. Whisper's open release democratized this capability, enabling voice interfaces and transcription in applications where cost or privacy previously prevented it.
Understanding Whisper helps organizations evaluate build-vs-buy decisions for audio transcription needs.
Examples in Practice
A podcast company uses Whisper to automatically transcribe every episode for SEO and accessibility, a task that would have cost thousands monthly with human transcription.
A developer builds a local voice note app with Whisper, ensuring personal recordings never leave the user's device while maintaining professional-grade accuracy.