This is a NextJS-based demo of AI.JSX, a JavaScript framework for building applications with Large Language Models. For more info, check out the AI.JSX docs, or the source code for this app.

This demo exercises several real-time ASR (speech-to-text) implementations. You can see how they do on a stock text recording using Start File, or you can use Start Mic to try with your own voice.

Deepgram

Latency: 0 ms
WER: 0.000

AssemblyAI

Latency: 0 ms
WER: 0.000

Speechmatics

Latency: 0 ms
WER: 0.000

Rev AI

Cost: $0.02/min
Latency: 0 ms
WER: 0.000

Soniox

Latency: 0 ms
WER: 0.000

Gladia

Latency: 0 ms
WER: 0.000