12 comments

[ 3.4 ms ] story [ 40.1 ms ] thread
I like these projects but they all have one flaw: the "open source" part refers only to the easy frontend, not to the backend doing the heavy lifting. With time, chances of CMUSphinx succeeding are getting lower because everybody is using proprietary APIs.
That's a great idea as for the next engine to add. The plan is to have multiple ones and not only Google's Speech API.
Voice to Text via https://speechtyping.com/voice-to-text-english

I have bookmarked this a while ago. May not be open source, and does not use the internet to convert. Would work for many use cases, but wont be as good as Google's AI.

works with just the browser with help of webkitSpeechRecognition built into the browser. Works on both Chrome and Safari (not Firefox).

If anyone is interested, Mozilla’s Deepspeech (https://github.com/mozilla/DeepSpeech) is also quite a solid open source offline alternative, and runs quite well even on low-end hardware like a Raspberry Pi.
Agreed, though what I was mainly missing with DeepSpeech was Speaker Diarization, which Google Speech API supports.
(comment deleted)