Home / Tools / Whisper

Whisper

AI transcription

Visit Whisper
Whisper preview
github.com Free

Upvotes

0

Pricing

Free

Category

Audio Tools

In the database since

Oct 2026

Overview

About Whisper

Whisper is OpenAI's open-source speech recognition model. It turns spoken audio into text accurately. Trained on 680,000 hours of audio from the web, it handles accents, background noise, and technical words well. It supports about 99 languages and can translate other languages into English. The model runs on your own computer, so audio never leaves your machine. You use it from the command line. Developers add it to apps for transcription, subtitles, and voice features. Anyone can download it, since the code and weights are public.

Key Features

Speech Recognition

Turns spoken audio into accurate written text. It works on podcasts, meetings, interviews, and videos with high quality.

99 Languages

Supports about 99 languages. It detects the spoken language on its own, then writes the transcript in that language.

Speech Translation

Translates speech from other languages into English. Useful when you need to understand audio you cannot speak.

Runs Locally

The model runs on your own computer. Audio never uploads anywhere, which keeps private recordings fully private.

Model Sizes

Comes in sizes from tiny to large. Smaller models run fast on weak hardware; larger ones give better accuracy.

Frequently Asked Questions (FAQ)

What is OpenAI Whisper?

Whisper is an open-source speech recognition model from OpenAI. It converts spoken audio into written text and supports about 99 languages.

Is Whisper an app I can download?

Whisper is a model, not a finished app. It has no buttons or menus — you run it from the command line, or use a wrapper app like Buzz that puts a friendly screen on top of it.

Which languages does Whisper support?

About 99. OpenAI trained it on audio in nearly a hundred languages, and it can also translate speech from those languages into English.

Can Whisper translate speech?

Yes. Set it to translation mode and it writes out what was said in English, even when the speaker used another language.

Do I need the internet to use Whisper?

No. Once you download the model files, everything runs on your own machine. Your audio never leaves your computer.

Can developers use Whisper in their apps?

Yes. OpenAI released the code and model weights publicly, so developers can add transcription to their own software.

Conclusion

Whisper gives you research-grade speech recognition without a subscription. Nearly a hundred languages, English translation, and local running make it a rare combination. It suits developers and anyone who needs private transcription. There is no app to download — it is a model for people who are comfortable with the command line or apps built on it.

Newsletter

Tools like this change fast

New Audio Tools tools launch every week — the newsletter keeps you ahead of what's worth trying.

Keep exploring

Similar AI tools

Ranked by the job they do, their capabilities, and how closely they fit this workflow.