Maybe the second-best-thing to come out of OpenAI after ChatGPT. Whisper is an open-source program that has revolutionized speech-to-text. Now, anyone has access to high-quality, high-speed transcription directly on their phone or computer.
Programmers have had a field day with this technology. I find myself using speech-to-text all the time, in programs ranging from [[Granophone]] to experimental tools like [[Voicebox]].
# Using Whisper in a Python Application
The `openai-whisper` package makes it incredibly easy to start integrating Whisper into an application:
1. Install globally via pip3
`pip3 install openai-whisper`
2. You should now be able to call whisper directly on an audio file
`whisper audio.mp3`
3. You can also call whisper via python
```
import whisper
model = whisper.load_model('base')
result = model.transcribe("file.mp3")
print(result['text'])
```