Using a CDN to bring in vosk-browser is the quickest way to integrate, broken down into the following steps:
- Introducing Scripts: In the HTML file's<head>maybe<body>Add a CDN link to the
<script src=’https://cdn.jsdelivr.net/npm/vosk-browser@0.0.5/dist/vosk.js’></script> - Loading Models: Download the .tar.gz zip for the corresponding language from the official model repository (e.g. https://ccoreilly.github.io/vosk-browser/models/)
- Initialization::
1. Asynchronous loading of model files through Vosk.createModel()
2. Create AudioContext and recognizer instances
3. Call getUserMedia to get microphone privileges.
4. Create audio data transfer pipeline through createTransferer
Note: The current latest version is 0.0.8, it is recommended to check for updates via jsDelivr. Small English models (e.g. vosk-model-small-en-us) are about 50MB, suitable for quick testing.
This answer comes from the articleVosk-Browser: Speech Recognition Tool Running in a BrowserThe
May not be reproduced without permission:AI productivity tools " How to quickly integrate vosk-browser into a web project via CDN?
English 
简体中文 
日本語 
Deutsch 
Português do Brasil