Skip to main content

​
Cartesia STT

Streaming speech-to-text using Cartesia’s ink-whisper model. Ported from LiveKit Agents (Apache 2.0) — uses ws, no vendor SDK required.

​
Quickstart

import { CartesiaSTT } from "getpatter";

const stt = new CartesiaSTT("csk_...", { language: "en" });
stt.onTranscript((t) => console.log(t.text, t.isFinal, t.confidence));
await stt.connect();
stt.sendAudio(pcmBuffer); // 16 kHz PCM s16le
// ...
stt.close();
Supported sample rates: 8000, 16000, 24000, 44100, 48000 Hz.