Reputation: 5347
I want to create a robot with Node.js and Tessel.io, i bought the audio module and i'm wondering if it exist any speech to text or audio to text package for node js or something else.
Can someone help me with this?
Thank you
Upvotes: 17
Views: 19213
Reputation: 111
I use vosk. It's a python library for voice recognition in multiple language. There is a wrapper for node js! It's great for offline recognition. Only must be download the model here (https://alphacephei.com/vosk/models) and then send the path of the unzipped model in your code like this demo (MODEL_PATH): https://github.com/alphacep/vosk-api/blob/master/nodejs/demo/test_microphone.js
I tried it in node js 16 and it works great Happy code! ;)
Upvotes: 2
Reputation: 104
You can try Picovoice. Depending on what you want the robot to do you can use the wake word, speech-to-intent, or speech-to-text.
Upvotes: 1
Reputation: 677
This issue is a bit stale, so I thought I'd add a more up-to-date answer:
CMUSphinx is great if you have something powerful enough to run it, but if you are on an embedded device like the Pi or the CHIP it can be quite slow.
One alternative is Sonus which has offline hotword detection and then uses Google Cloud Speech for streaming results.
Upvotes: 3
Reputation: 25210
There is a node module for CMUSphinx
https://github.com/cmusphinx/node-pocketsphinx
It works offline. It's a bit initial state but let me know if you have questions on it.
Upvotes: 8