Whisper
OpenAI's open-source speech-to-text model that converts audio to text with high accuracy across 99 languages.
What is Whisper?
Whisper is OpenAI's open-source automatic speech recognition (ASR) model that transcribes audio into text.
It's trained on 680,000 hours of multilingual audio from the web, which means it handles accents, background noise, and technical language better than most transcription tools.
Builders use it through the OpenAI API, run it locally via Python, or integrate it into apps for real-time transcription. It works for both transcription (audio to text in the same language) and translation (audio to English text).
The API costs $0.006 per minute of audio. The model is fully open-source on GitHub, so you can run it locally for free if you have the compute.
Good to Know
Trained on 680,000 hours of multilingual audio data
Supports 99 languages for transcription and translation to English
Handles noisy audio, accents, and technical terminology well
Available via API ($0.006/minute) or open-source for local use
Built on transformer architecture with encoder-decoder design
How Vibe Coders Use Whisper
Transcribing customer support calls to build a searchable knowledge base
Adding automatic captions to your product demo videos
Building a voice interface for your app that understands multiple languages
Transcribing podcast episodes to generate show notes and blog posts
Frequently Asked Questions
Related Terms
Computer systems that learn from data and perform tasks that typically require human intelligence, like recognizing patterns and making decisions.
AI technology that converts spoken words into written text in real-time or from recordings.
AI voice platform that generates ultra-realistic speech from text, clones voices, and dubs content into 29+ languages.
The AI research company behind ChatGPT, GPT-4, DALL-E, and the APIs that power thousands of AI products.
Technology that converts written text into spoken audio using AI-generated voices that sound increasingly human-like.
Join 0 others building with AI