Whisper: open-source AI that turns speech into text

Whisper is a speech-to-text recognition model released as open source, and it quickly became a popular choice for transcription.

What can Whisper do?

• Multilingual: recognizes and even translates many languages, including Vietnamese.
• Runs offline: can run on a personal machine without sending audio over the network.
• Multiple sizes: from a small model for laptops to a large one for high accuracy.
• Strong community: accelerated variants like whisper.cpp make it run smoother.

Real-world uses

For recruiters or content creators, Whisper is genuinely useful: transcribe interviews into text for records, generate subtitles for company videos, or auto-note meetings — all with a free model you can self-host.