How to use it in react js , with custom wake word?

[OnlinePage]

Hi i am looking forward for good hotword detection library purely in node js, and i found WebVoiceSDK, it's performance in the demo is really good.
Now can you tell me how to utilise it in react js and with custom wakeword?

any refrence ? if so then this library has very good potenital to scale up.

[damienlaine]

Greetings.
It's a front-end package. Part of linto.ai ecosystem.

How to use it in React --> Import script in your HTML page and bind some events to your workflow. Look examples in source code.

To train custom HotWord --> https://doc.linto.ai/docs/developpers/agent/clients/custom_hotwords It's something that requires a bit of expertise. We can help on this. It requires audio data (thousands of hours of audio withtout WakeWord) and around 200 recordings (multiple locutors) of the WakeWord you want to train.

You might also want to dig in https://github.com/linto-ai/linto-web-client

Final words : This tool powers LinTO web client and has been recently used in industry for PWAs like Hello Casto system https://www.castorama.fr/hello-casto

[OnlinePage]

Greetings.
It's a front-end package. Part of linto.ai ecosystem.

How to use it in React --> Import script in your HTML page and bind some events to your workflow. Look examples in source code.

To train custom HotWord --> https://doc.linto.ai/docs/developpers/agent/clients/custom_hotwords It's something that requires a bit of expertise. We can help on this. It requires audio data (thousands of hours of audio withtout WakeWord) and around 200 recordings (multiple locutors) of the WakeWord you want to train.

You might also want to dig in https://github.com/linto-ai/linto-web-client

Final words : This tool powers LinTO web client and has been recently used in industry for PWAs like Hello Casto system https://www.castorama.fr/hello-casto

Awesome , going to try the implementation!!

Wouldn't be much better , if there's an npm module for it 🤔. Also easy to use custom wake word trainer with ui. 😀

Or we can create onnxx runtime compatible models for wake word detection using pytorch or tflite which would be much lighter.

[damienlaine]

Wouldn't be much better , if there's an npm module for it -->

It's a static JS script. intended for "in browser" voice operation. What's more easy than a <script> tag ?
Btw, you can always install this front package using NPM https://www.npmjs.com/package/@linto-ai/webvoicesdk

We use tflite models in the android / flutter version of the LinTO web client.

[OnlinePage]

Wouldn't be much better , if there's an npm module for it -->

It's a static JS script. intended for "in browser" voice operation. What's more easy than a <script> tag ?
Btw, you can always install this front package using NPM https://www.npmjs.com/package/@linto-ai/webvoicesdk

We use tflite models in the android / flutter version of the LinTO web client.

Oo cool! So what do you suggest for the ionic capacitor app?