Voice Dictation Setup Guide for Mac and Windows
WisperCode Team · February 4, 2026 · 15 min read
TL;DR
WisperCode is a voice dictation app that runs entirely on your machine using OpenAI's Whisper model. It works on both macOS and Windows with no cloud processing and no account required. Installation takes under five minutes, and this guide walks you through every step from downloading the app to optimizing your setup for the best transcription accuracy.
What You Need Before Starting
Before you download WisperCode, make sure your system meets the minimum requirements.
| Requirement | macOS | Windows |
|---|---|---|
| Operating System | macOS 12 Monterey or later | Windows 10 or Windows 11 |
| Processor | Apple Silicon (M1/M2/M3/M4) or Intel | x64 processor (Intel or AMD) |
| RAM | 4 GB minimum, 8 GB recommended | 4 GB minimum, 8 GB recommended |
| Disk Space | 500 MB for app + model files | 500 MB for app + model files |
| Microphone | Built-in or external | Built-in or external |
| Internet | Required for initial model download only | Required for initial model download only |
A built-in laptop microphone will work for basic dictation. For professional use or noisy environments, an external microphone is recommended. After the initial Whisper model download, WisperCode works entirely offline.
Step 1: Download WisperCode
Visit the download page and select the installer for your operating system. WisperCode is free during the beta period -- no payment information or account creation required.
For macOS, you will download a .dmg disk image file. For Windows, you will download a .exe installer. Both files are typically around 50 MB. The Whisper model itself is downloaded separately on first launch, so the initial installer download is fast.
Step 2: Install on macOS
Follow these steps to install WisperCode on your Mac.
-
Open the .dmg file. Double-click the downloaded
.dmgfile. A Finder window will open showing the WisperCode app icon and a shortcut to your Applications folder. -
Drag to Applications. Drag the WisperCode icon onto the Applications folder shortcut. This copies the app to your Applications directory.
-
Handle the Gatekeeper warning. When you first launch WisperCode, macOS will display a warning that the app is from an unidentified developer. This is normal for apps distributed outside the Mac App Store. To proceed:
- Click "Cancel" on the initial warning dialog.
- Open System Settings and navigate to Privacy & Security.
- Scroll down to the Security section. You will see a message about WisperCode being blocked.
- Click Open Anyway and confirm.
- Alternatively, right-click the app in Finder, select "Open," and click "Open" in the dialog.
-
Launch the app. After clearing the Gatekeeper warning, WisperCode will launch and appear as "Wisper Code" in your menu bar at the top of the screen. You will see a small icon in the menu bar tray area, confirming the app is running.
-
Complete onboarding. On first launch, a brief setup wizard walks you through granting permissions and choosing your initial settings. Follow the prompts to get started quickly.
Step 3: Install on Windows
Follow these steps to install WisperCode on Windows.
-
Run the installer. Double-click the downloaded
.exefile. If Windows SmartScreen displays a warning, click More info and then Run anyway. This warning appears because WisperCode is a newer application and has not yet built up a reputation with Microsoft's SmartScreen filter. -
Follow the installation wizard. The installer will guide you through the process. The default install location is
C:\Program Files\WisperCode. Unless you have a specific reason to change it, the default location is fine. -
Choose startup options. During installation, you can choose whether WisperCode should launch automatically when Windows starts. If you plan to use voice dictation regularly, enabling this option saves you from manually starting the app each time.
-
Launch WisperCode. After installation completes, launch WisperCode from the Start menu or desktop shortcut. The app will appear in the Windows system tray (the small icons area in the taskbar near the clock).
-
Complete onboarding. The first-launch wizard will guide you through permissions and initial configuration, just as it does on macOS.
Step 4: Grant Permissions
WisperCode needs specific permissions to function. Without these, the app cannot record your voice or type transcribed text into other applications.
macOS Permissions
Microphone Access
WisperCode needs microphone access to hear your voice. macOS will prompt you automatically on first launch.
- If you missed the prompt, go to System Settings then Privacy & Security then Microphone.
- Find WisperCode in the list and toggle it on.
- You may need to restart WisperCode after granting this permission.
Why it is needed: Without microphone access, WisperCode literally cannot hear you speak. The audio is processed entirely on your machine and never transmitted anywhere.
Accessibility Access
WisperCode needs accessibility permission to simulate keyboard input and type transcribed text into the active application.
- Go to System Settings then Privacy & Security then Accessibility.
- Click the lock icon to make changes (you may need to enter your password).
- Find WisperCode in the list and toggle it on, or click the plus button to add it manually.
- You will need to restart WisperCode after granting this permission.
Why it is needed: macOS restricts which apps can simulate keystrokes for security reasons. Without accessibility permission, WisperCode can transcribe your speech but cannot type the result into other applications.
Windows Permissions
Microphone Access
Windows will prompt for microphone access on first use.
- If you need to grant it manually, go to Settings then Privacy then Microphone.
- Make sure "Allow apps to access your microphone" is turned on.
- Scroll down and ensure WisperCode is listed and enabled.
Why it is needed: Same as macOS. WisperCode needs to capture audio from your microphone for transcription.
Keyboard and Accessibility
On Windows, the ability to simulate keyboard input and register global hotkeys is handled by the installer. No additional manual configuration is needed. If you installed WisperCode using the standard installer, these capabilities are already set up.
Step 5: Choose Your Whisper Model
WisperCode uses OpenAI's Whisper model for speech recognition. Multiple model sizes are available, and choosing the right one depends on your hardware and accuracy needs. You can change the model at any time in Settings.
| Model | Size | RAM Needed | Speed | Accuracy | Best For |
|---|---|---|---|---|---|
| Tiny | ~75 MB | 1 GB | Fastest | Basic | Older hardware, quick notes |
| Base | ~150 MB | 1 GB | Fast | Good | Most users starting out |
| Small | ~500 MB | 2 GB+ | Moderate | Better | Daily dictation with decent hardware |
| Medium | ~1.5 GB | 5 GB+ | Slower | Very Good | Professional use, complex vocabulary |
| Large-v3 | ~3 GB | 10 GB+ | Slowest | Best | Maximum accuracy, GPU recommended |
Our recommendation: Start with the Base model. It offers a good balance of speed and accuracy, runs well on most hardware, and downloads quickly. If you find that it misrecognizes words frequently, move up to Small or Medium. If you have a powerful machine with plenty of RAM (or a dedicated GPU), the Large-v3 model provides the best accuracy available.
The model downloads automatically on first launch. You can switch models later in Settings without reinstalling the app. Each model downloads only once and is cached locally.
For a deeper comparison of how each model performs, see Whisper Model Sizes Compared. To learn more about the underlying technology, read What Is OpenAI Whisper.
Step 6: Configure Your Hotkey
WisperCode uses a global hotkey to start and stop recording. The default is Ctrl+Space (or Cmd+Space on macOS if not in conflict with Spotlight). You can change this to any key combination in Settings.
There are four hotkey modes, each suited to different workflows.
Hold (Default)
Hold the hotkey to record. Release it to stop recording and trigger transcription. This is the most intuitive mode and the one we recommend for beginners. It mirrors the push-to-talk pattern used in voice chat applications.
Toggle
Press the hotkey once to start recording. Press it again to stop. This is useful for longer dictation sessions where holding a key would be uncomfortable. You can speak for as long as you need without keeping a key pressed.
Press
Press the hotkey once to start recording. WisperCode automatically stops recording when it detects a sustained pause in your speech. This is a hands-free option that works well for short dictation bursts.
Double-Press
Double-tap the hotkey to start recording. A single press stops it. This mode reduces accidental activation since a deliberate double-tap is less likely to happen by accident than a single press.
To change your hotkey or mode, open WisperCode Settings and navigate to the Hotkey section. You can set any key combination that does not conflict with your operating system or other applications.
Step 7: Test Your First Dictation
Now that everything is configured, test your setup.
-
Open any text application. A text editor, word processor, email client, browser text field, or code editor all work. WisperCode types into whatever application is currently focused.
-
Click where you want text to appear. Place your cursor in the text field where you want the transcribed text inserted.
-
Press your hotkey. If you are using Hold mode (the default), press and hold Ctrl+Space. You should see a recording indicator appear, confirming that WisperCode is listening.
-
Speak a test sentence. Try something simple like "The quick brown fox jumps over the lazy dog." Speak at a normal pace and volume. You do not need to speak slowly or loudly.
-
Release the hotkey. After a brief moment, the transcribed text will appear at your cursor position.
If Nothing Happens
If text does not appear after releasing the hotkey, check these items in order:
- Permissions: Verify that both microphone and accessibility permissions are granted (Step 4).
- Microphone: Make sure your microphone is connected and selected as the input device in your system sound settings.
- Hotkey conflict: Another application may be using the same key combination. Try changing the hotkey in WisperCode Settings.
- App is running: Confirm WisperCode is running by looking for its icon in the menu bar (macOS) or system tray (Windows).
- Volume: Speak closer to the microphone or increase the input volume in your system settings.
Optimizing Your Setup
Once basic dictation is working, these adjustments will improve your experience.
Microphone Tips
Your microphone has a significant impact on transcription accuracy. A few practical adjustments make a noticeable difference.
- Distance matters. Position yourself 6 to 12 inches from the microphone. Too close introduces plosive sounds (harsh "p" and "b" noises). Too far picks up room noise.
- Reduce background noise. Close windows, turn off fans if possible, and avoid dictating in crowded spaces. Whisper handles some background noise well, but a quiet environment always produces better results.
- Use a directional microphone. If you work in a shared space, a cardioid microphone rejects sound from the sides and back, focusing on your voice. USB podcast microphones and headset microphones both work well.
- Check your input level. In your system sound settings, make sure your microphone input level is set so that normal speech registers in the middle of the volume meter, not peaking at the top.
For specific product recommendations, see Best Microphones for Voice Dictation.
Adding Vocabulary Hints
If you regularly use technical terms, proper nouns, or industry jargon, vocabulary hints help Whisper recognize them correctly. Open WisperCode Settings and navigate to the Vocabulary section. Add terms like framework names, medical terminology, legal phrases, or company-specific words.
For example, adding "Kubernetes," "PostgreSQL," or "HIPAA" as vocabulary hints significantly improves recognition of these terms during dictation.
See Vocabulary Hints for Technical Terms for a detailed walkthrough.
Filler Word Removal
Natural speech includes filler words like "um," "uh," "like," and "you know." WisperCode can automatically strip these from your transcriptions so the output reads cleanly. Enable this feature in Settings under the Text Processing section.
Learn more in Filler Word Removal for Cleaner Text.
Context-Aware Styling
WisperCode can detect which application you are dictating into and adjust text formatting accordingly. Dictating into a chat app might produce more casual, shorter sentences. Dictating into a document editor might produce more formal, properly punctuated text. Configure style profiles in Settings under the Styling section.
For full details, see the Context-Aware Styling Guide.
Snippets and Text Expansion
Snippets let you define short trigger phrases that expand into longer text blocks. For example, you could set up a snippet where saying "email signature" inserts your full name, title, and contact information. This is especially powerful for repetitive text you dictate frequently.
Set up snippets in Settings under the Snippets section, or read the Voice Snippets and Text Expansion Guide.
Troubleshooting Common Issues
| Issue | Possible Cause | Solution |
|---|---|---|
| No audio detected | Microphone not connected or permission not granted | Check microphone connection, verify permissions in system settings, ensure correct input device is selected |
| Hotkey not working | Key conflict with another app, or accessibility permission missing | Change hotkey in Settings, verify accessibility permission (macOS), try a different key combination |
| Slow transcription | Large model on limited hardware | Switch to a smaller model (Base or Tiny), close other resource-intensive applications |
| Wrong words in output | Background noise, unclear speech, or missing vocabulary hints | Improve microphone setup, add vocabulary hints for specialized terms, try a larger model |
| App not starting | Installation incomplete or OS blocking the app | Reinstall the app, check Gatekeeper (macOS) or SmartScreen (Windows) settings, run as administrator on Windows |
| Text not appearing in target app | Accessibility permission not granted, or target app blocking input | Grant accessibility permission (macOS), try a different target application, check if the target app is running in admin mode (Windows) |
Platform-Specific Notes
macOS Notes
- Menu bar icon. WisperCode runs as a menu bar app. The icon appears in the top-right area of your screen alongside other menu bar icons. If you do not see it, check if it is hidden behind the notch on newer MacBooks (you can rearrange menu bar icons by holding Cmd and dragging).
- Accessibility re-prompt. After macOS updates, the system may revoke accessibility permissions. If WisperCode stops typing text after an OS update, revisit System Settings and re-enable accessibility access.
- Apple Silicon performance. WisperCode runs natively on Apple Silicon. M-series chips handle Whisper models efficiently, and you can comfortably use the Small or Medium model on machines with 8 GB of RAM or more.
Windows Notes
- Administrator mode applications. If you are trying to dictate into an application running as administrator (such as an elevated Command Prompt or some IDE configurations), WisperCode may not be able to type into it unless WisperCode is also running with elevated privileges. Right-click the WisperCode shortcut and select "Run as administrator" if you encounter this issue.
- Antivirus false positives. Some antivirus software may flag WisperCode because it registers global hotkeys and simulates keyboard input. These are legitimate functions required for dictation. If your antivirus blocks WisperCode, add it to the exception list.
- Startup behavior. If you enabled the "Start with Windows" option during installation, WisperCode will appear in the system tray after each login. You can change this later in Settings or via the Windows Startup Apps settings page.
Frequently Asked Questions
Does WisperCode work offline?
Yes. After the initial Whisper model download (which requires an internet connection), WisperCode works entirely offline. All speech processing happens locally on your machine. You can dictate on a plane, in a cabin with no Wi-Fi, or on a secured network with no internet access.
Can I use WisperCode with any application?
Yes. WisperCode types into whatever application currently has focus. This includes text editors, word processors, email clients, web browsers, chat applications, code editors, terminal windows, and any other application that accepts keyboard input. You place your cursor where you want text to appear, press the hotkey, speak, and the transcription is inserted at the cursor position.
Which Whisper model should I start with?
Start with the Base model. It offers a strong balance of speed and accuracy, works on virtually all hardware, and downloads quickly. If you find accuracy lacking for your use case (particularly with specialized vocabulary or accented speech), step up to Small or Medium. Only use the Large-v3 model if you have 10 GB or more of available RAM and prioritize maximum accuracy over speed.
How do I change the hotkey?
Open WisperCode Settings (click the menu bar or system tray icon and select Settings). Navigate to the Hotkey section. From there, you can change the key combination and switch between Hold, Toggle, Press, and Double-press modes. The new hotkey takes effect immediately.
Is WisperCode free?
Yes. WisperCode is free during the beta period. There is no payment information required, no account to create, and no feature limitations. Download it from the download page and start using it immediately.
Try WisperCode free during beta -- Download
Related Articles
Filler Word Removal: Get Cleaner Transcriptions
Learn how automatic filler word removal works in voice dictation. Remove ums, uhs, and false starts from your transcriptions without manual editing.
January 22, 2026 · 9 min read
Best Microphones for Voice Dictation in 2026
Find the best microphone for voice dictation. Compare USB mics, headsets, and lapel mics across price, noise rejection, and dictation accuracy.
January 12, 2026 · 15 min read
Privacy-First Voice Dictation: The Complete Guide
Learn how local voice dictation protects your data. Compare cloud vs on-device speech recognition for privacy, security, and compliance.
February 5, 2026 · 15 min read