Home/AI/Whisper
W
Capabilities:audio
License:MIT
Open sourceAPI availableLocal runtimeOpen weights

Whisper

About

OpenAI's open-source speech recognition model that transcribes audio in 99 languages with near-human accuracy. Runs locally via CLI, Python or the OpenAI API.

Pricing

Free

Whisper is an automatic speech recognition (ASR) model that turns audio into text with robust multilingual support. It was trained on 680,000 hours of web audio and handles accents, noise and technical terms well.

Key Features

  • 99 languages with translation to English.
  • Local inference — no cloud required.
  • Model sizes from tiny to large for any hardware.
  • Word-level timestamps.
  • CLI, Python API and OpenAI API.
  • Fine-tunable for domain-specific audio.

Pricing

Whisper is free and open source under the MIT license. The hosted API version costs ~$0.006/minute.

Getting Started

pip install -U openai-whisper
whisper audio.mp3 --model small --language en

For the API: curl https://api.openai.com/v1/audio/transcriptions -F [email protected] -F model=whisper-1