dictate.js is a small Javascript library for browser-based real-time speech recognition. It uses Recorderjs for audio capture, and a WebSocket connection to the Kaldi GStreamer server for (Estonian and English) speech recognition.

The API is modelled after Android's SpeechRecognizer. See the source code of lib/dictate.js and the usage in demos/demo.js.

To run the demos on localhost, start a local webservice, e.g.:

python -m SimpleHTTPServer

and then open e.g. http://localhost:8000/demos/diff.html.

In order to avoid the repeated request for access to the microphone use a https-URL, this works e.g. with Google App Engine.

Some demos are available here.

Known to work in

• Chromium 31.0.1650.63 on Ubuntu desktop
• Chrome Beta 32.0.1700.94 on Android