How to Dictate Code with Murmur: Complete Guide

TL;DR

Dictating code with Murmur is simple: press Ctrl+Space, speak your requirement, release, and let AI tools (Cursor, Claude Code) generate the code. The key is using natural language instead of trying to dictate exact syntax. Learn the technique in 5 minutes and start coding hands-free today.

Why Dictate Code?

Before diving into how, let's understand why you'd want to dictate code:

Speed — Speaking a requirement is faster than typing it, especially for complex prompts. "Create a React component that fetches users from an API and displays them in a table with loading and error states" takes 10 seconds to speak but 45 seconds to type.

Accuracy — AI tools like Cursor and Claude Code respond better to conversational, detailed prompts. When you dictate naturally, you tend to be more thorough and explicit about requirements.

Health — Reduce keyboard strain and repetitive stress injuries. Your hands get a break while your voice does the work.

Focus — Stay focused on the problem you're solving instead of worrying about typing mechanics.

Setup: 5-Minute Installation

Step 1: Download Murmur

Visit Murmur's download page and grab the Windows or macOS version.

Step 2: Install

• Windows — Run the installer, follow prompts, grant microphone permission
• macOS — Drag Murmur to Applications folder, grant microphone and accessibility permissions

Step 3: Configure Your Shortcut

The default shortcut is Ctrl+Space (Windows) or Cmd+Space (macOS). This works in almost all editors.

If it conflicts with your editor's autocomplete, customize it in Murmur settings:

• Open Murmur settings
• Navigate to "Keyboard Shortcuts"
• Choose your preferred activation key (Alt+`, Ctrl+M, etc.)

Step 4: Test It

1. Open VS Code, Cursor, or any text editor
2. Click in a text field
3. Press Ctrl+Space
4. Say "Hello world"
5. Release the key
6. You should see "Hello world" typed in your editor

Congratulations! You're ready to dictate code.

The Right Technique: Speak Natural Language, Not Syntax

The biggest mistake beginners make is trying to dictate exact code syntax:

❌ Wrong: "Type def function name add numbers colon new line...indent return..."

✓ Right: "Create a function that adds two numbers and returns the result"

When dictating code, your job is to describe what you want, not to transcribe syntax. Let AI tools handle the syntax.

Example 1: React Component

Bad approach (dictating syntax): "Const function my component with props equals curly braces..."

Good approach (natural language): "Create a React component called UserCard that accepts user props and displays their name, email, and avatar"

Then let Claude Code or Cursor generate the component.

Example 2: API Request

Bad approach: "Fetch from URL then JSON then..."

Good approach: "Create a function that fetches user data from slash api slash users endpoint and handles loading and error states"

AI tools will generate clean, idiomatic code based on your description.

Dictating with Cursor

Cursor is purpose-built for AI-assisted development, making it perfect for voice dictation.

Workflow

1. Open Cursor and your project
2. Open Cursor Chat (Cmd+K or Ctrl+K)
3. Press Ctrl+Space to activate Murmur
4. Speak your requirement clearly and naturally
5. Release and Murmur transcribes your voice
6. Press Enter and Cursor generates code
7. Review and iterate using voice or keyboard

Real Example

You want a function that parses CSV data.

What you say: "I need a function that parses CSV data and returns an array of objects with headers as keys"

Cursor generates:

function parseCSV(csvText) {
  const lines = csvText.trim().split('\n');
  const headers = lines[0].split(',');
  const data = lines.slice(1).map(line => {
    const values = line.split(',');
    return headers.reduce((obj, header, i) => {
      obj[header] = values[i];
      return obj;
    }, {});
  });
  return data;
}

Time to dictate: 8 seconds Time to type: 45 seconds Saved: 37 seconds

Multiply this across your day and you'll save hours.

Dictating with Claude Code

Claude Code (via Claude.ai) is excellent for voice-first development.

Workflow

1. Open Claude.ai and start a new chat
2. Open Claude Code (bottom of screen)
3. Press Ctrl+Space to activate Murmur
4. Describe your feature or task
5. Let Claude implement it in real-time
6. Iterate by voice if needed

Real Example

"Build me a to-do app in React with ability to add, delete, and mark tasks as done. Use React hooks and store data in localStorage"

Claude Code will generate a complete, working app in your browser. No typing required.

Dictating with VS Code + Copilot

VS Code with GitHub Copilot Chat works similarly.

Workflow

1. Open VS Code
2. Open Copilot Chat (Cmd+Shift+I or Ctrl+Shift+I)
3. Press Ctrl+Space to activate Murmur
4. Describe your task
5. Copilot generates code

The key advantage: you're working in your actual development environment, not a browser.

Pro Tips for Better Voice Dictation

1. Use Context Markers

Help Murmur understand what you're dictating by being explicit about language and context:

✓ "Python function to validate email addresses" ✓ "JavaScript async function that fetches data" ✓ "SQL query to find users created in the last week"

2. Spell Out Technical Terms

For uncommon variable names or domain-specific terms, spell them out:

"Create a function called S-T-L-C mapper that..."

This ensures Murmur transcribes correctly instead of guessing.

3. Describe the Interface, Not Implementation

Tell AI what the function should do, not how to do it:

✓ "Create a function that validates passwords: 8+ characters, at least one number, one uppercase letter"

✗ "Create a function that checks length greater than 8 and..."

The AI will pick the right approach.

4. Break Complexity into Chunks

If your requirement is complex, break it into multiple smaller prompts:

Instead of: "Create a complete authentication system with login, registration, password reset, and two-factor authentication"

Try: "Create a login function", then "Add registration logic", then "Add password reset flow"

This gets you better results faster.

5. Use Natural Grammar

Murmur is designed to understand natural speech. Speak like you're explaining to a colleague:

✓ "I need to sort this array by the user's creation date in descending order"

✗ "Sort array creation date descending"

Natural language gets transcribed more accurately.

Dictation Dos and Don'ts

Do

✓ Use natural language and complete sentences ✓ Include context about your requirements ✓ Spell out proper nouns and technical terms ✓ Review AI-generated code before accepting it ✓ Iterate iteratively with multiple prompts ✓ Take breaks to rest your voice and review progress

Don't

✗ Try to dictate exact syntax ✗ Rush through your explanation ✗ Use acronyms without spelling them out ✗ Give vague requirements ("make it work") ✗ Dictate more than 2-3 minutes at once ✗ Rely 100% on AI—always review and test code

Real-World Examples

Example 1: Building an API Endpoint

What you say: "Build an Express endpoint at slash API slash products that accepts a GET request and returns all products from a MongoDB database, with error handling"

AI generates: Complete working endpoint with proper error handling

Result: 30 seconds of dictation vs 10 minutes of typing

Example 2: React Hook

What you say: "Create a custom React hook called useLocalStorage that lets me read and write to browser local storage with automatic sync"

AI generates: Full hook implementation with proper cleanup

Result: Voice dictation faster and cleaner than manual coding

Example 3: Utility Function

What you say: "I need a utility function that converts snake case variable names to camel case"

AI generates: Properly tested utility function

Result: 5 seconds of dictation, immediate working code

Integrating Voice Dictation into Your Workflow

Morning Setup

1. Open your IDE and Murmur
2. Ensure your microphone is working (test with Murmur's test feature)
3. Review your task list for the day

During Coding

• Use voice for:

  • Long prompts to AI tools
  • Feature descriptions
  • Bug reports and documentation

• Use keyboard for:

  • Quick edits
  • Debugging
  • Manual refactoring

Context Switching

When switching between different coding tasks, take a 2-minute break for your voice. Continuous dictation for more than 20 minutes can lead to fatigue.

Troubleshooting Common Issues

Murmur Doesn't Hear Me

• Check microphone permissions in system settings
• Ensure your microphone is not muted
• Try moving closer to the microphone
• Test with Murmur's test feature first

Poor Transcription Accuracy

• Speak more clearly and slowly
• Eliminate background noise
• Use a better quality microphone (USB headsets recommended)
• Spell out technical terms

Shortcut Conflicts

If Ctrl+Space conflicts with your editor:

• Customize the shortcut in Murmur settings
• Common alternatives: Alt+`, Ctrl+M, Ctrl+Shift+V

The Future of Voice-First Development

Voice dictation is becoming table stakes for developer tools. Cursor, Claude Code, and Copilot all recognize the efficiency gains and are optimizing their interfaces for natural language input.

Adopting voice dictation now puts you ahead of the curve. You'll develop faster, with less strain, and with cleaner prompts that AI tools respond better to.

Ready to start dictating code? Download Murmur today and experience hands-free development. Start with the free tier—if you're dictating more than 5 times per day, you'll want the Pro plan.

Related articles:

• Murmur for Developers on Windows & macOS
• Voice Coding on Mac: VS Code + Murmur Setup Guide