murmr/Documentation
Try Playground

Quickstart

Generate your first audio in under 5 minutes. This guide covers installation, authentication, and making your first API call.

Prerequisites

  • 1.A murmr account — sign up free
  • 2.An API key — create one in your dashboard
  • 3.Node.js 18+ (for the SDK) or any language that supports HTTP

1. Install the SDK

npm install @murmr/sdk

No SDK? You can use the REST API directly with any HTTP client.

2. Initialize the Client

import { MurmrClient } from '@murmr/sdk';

const client = new MurmrClient({
  apiKey: process.env.MURMR_API_KEY!,
});

Keep your API key secret

Never expose your API key in client-side code or commit it to version control. Use environment variables.

3. Design a Voice and Generate Audio

Describe any voice in natural language and generate audio in a single call. This is murmr's core feature — no presets, no cloning, just describe what you want.

import { writeFileSync } from 'fs';

const audio = await client.voices.design({
  input: "Welcome to our app. Let me show you around.",
  voice_description: "A warm, friendly female voice, calm and clear",
  language: "English",
});

writeFileSync("welcome.wav", Buffer.from(audio));
console.log("Audio saved to welcome.wav");

4. Stream Audio (Low Latency)

For real-time playback, use the streaming endpoint. Audio chunks arrive via Server-Sent Events as they're generated — no waiting for the full file.

const response = await fetch("https://api.murmr.dev/v1/voices/design/stream", {
  method: "POST",
  headers: {
    "Authorization": `Bearer ${process.env.MURMR_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    text: "Hello! This audio is streaming to you in real time.",
    voice_description: "A confident male narrator voice",
    language: "English",
  }),
});

const reader = response.body!.getReader();
const decoder = new TextDecoder();

while (true) {
  const { done, value } = await reader.read();
  if (done) break;

  for (const line of decoder.decode(value).split("\n")) {
    if (!line.startsWith("data: ")) continue;
    const data = JSON.parse(line.slice(6));

    if (data.chunk) {
      // Base64-encoded PCM audio chunk (24kHz, 16-bit, mono)
      const pcm = Buffer.from(data.chunk, "base64");
      console.log(`Received ${pcm.length} bytes`);
    }

    if (data.done) {
      console.log(`Complete in ${data.total_time_ms}ms`);
      console.log(`Time to first chunk: ${data.first_chunk_latency_ms}ms`);
    }
  }
}

~450ms time-to-first-chunk

Streaming starts delivering audio in ~450ms. See the SSE Streaming guide for browser playback with Web Audio API.

5. Save a Voice for Reuse

Found a voice you like? Save it so every request sounds the same. Saved voices use voice embeddings — they're fast and consistent.

// Step 1: Design a voice and get the audio
const audio = await client.voices.design({
  input: "Testing this voice for my project.",
  voice_description: "A calm, professional narrator",
});

// Step 2: Save it via REST API
const response = await fetch("https://murmr.dev/api/v1/voices", {
  method: "POST",
  headers: {
    "Authorization": `Bearer ${process.env.MURMR_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    name: "Project Narrator",
    description: "Calm, professional narrator for product videos",
    audio: Buffer.from(audio).toString("base64"),
    language: "English",
  }),
});

const voice = await response.json();
console.log(`Saved as ${voice.id}`);
// e.g. "voice_abc123"

6. Generate with a Saved Voice

Once saved, use the voice ID in any speech request. Every call with the same voice ID produces the same voice.

import { MurmrClient, isSyncResponse } from '@murmr/sdk';
import { writeFileSync } from 'fs';

const result = await client.speech.create({
  input: "This is my saved voice. It sounds the same every time.",
  voice: "voice_abc123", // Your saved voice ID
  response_format: "mp3",
});

if (isSyncResponse(result)) {
  writeFileSync("output.mp3", Buffer.from(await result.arrayBuffer()));
}

Next Steps