In this demo, I am able to run SILERO VAD in the browser using the ONNX-web runtime. If you want to run voice activity detection in your Next.js app, you can copy the VAD directory and use the SpeechChunk class.
Remember to add the configuration present in the nextconfig.js to load the WASM files.
You can access the Demo here https://dictationdaddy.web.app/