@typestreamdev/recorder

Client SDK for Typestream — low-latency speech-to-text for developers.

Handles microphone capture, upload routing, and recording UI state. Your API key stays on the server.

Install

npm install @typestreamdev/recorder

For the styled React component: npm install framer-motion lucide-react

Quick start

"use client";

import { TypestreamRecorder } from "@typestreamdev/recorder";

export function VoiceInput() {
  return <TypestreamRecorder proxyUrl="/api/typestream" />;
}

Add a server-side proxy at /api/typestream that forwards to Typestream with your API key. See the docs.

Integration options

| Import | Use case | |--------|----------| | @typestreamdev/recorder | React hook + styled component | | @typestreamdev/recorder/vanilla | Framework-agnostic core | | @typestreamdev/recorder/web-component | <typestream-recorder> custom element |

// Headless hook — build your own UI
import { useTypestream } from "@typestreamdev/recorder";
const { state, transcript, start, stop, reset } =
  useTypestream({ proxyUrl: "/api/typestream" });

Next.js

// next.config.ts
transpilePackages: ["@typestreamdev/recorder"],

License

MIT