System Requirements
Vocoding runs natively on your Mac or PC. Check below to make sure your system is compatible before you get started.
macOS
Requirements for Mac computers
| Requirement | Minimum | Recommended |
|---|---|---|
| macOS version | macOS 13 Ventura | macOS 14 Sonoma or later |
| Processor | Apple M1 or Intel (2017 or newer) | Apple M1 or later |
| RAM | 4 GB | 8 GB or more |
| Disk space | 500 MB (app + base model) | 2 GB (for larger Whisper models) |
| Microphone | Any built-in or external mic | USB or built-in mic |
Apple Silicon (M1, M2, M3, M4)
Full support with GPU-accelerated transcription via Metal. All Whisper models work great. This is the best experience.
| Model | Performance | Use Case |
|---|---|---|
| Tiny (75 MB) | Instant | Quick commands, drafts |
| Base (142 MB) | Very fast | Daily use |
| Small (466 MB) | Fast | Best balance of speed and accuracy |
| Medium (1.5 GB) | Good | High accuracy needs |
| Large V3 Turbo (1.5 GB) | Good | Maximum accuracy, multilingual |
Intel Mac (2017-2020)
Supported with CPU-only transcription. No GPU acceleration. We recommend using smaller models for the best experience.
| Model | Performance | Recommendation |
|---|---|---|
| Tiny (75 MB) | Good | Recommended |
| Base (142 MB) | Moderate | Recommended |
| Small (466 MB) | Slow | Acceptable for short recordings |
| Medium or larger | Very slow | Not recommended |
Intel Macs run transcription on CPU only. For the fastest experience on Intel, use the Tiny or Base model. If you need higher accuracy, consider enabling cloud transcription (requires API key).
macOS Permissions Required
Vocoding needs the following permissions (you'll be prompted on first launch):
- 1Microphone — Required to record your voice
- 2Accessibility — Required for auto-paste functionality (optional but recommended)
Windows
Requirements for Windows PCs
| Requirement | Minimum | Recommended |
|---|---|---|
| Windows version | Windows 10 (version 1803+) | Windows 10 or 11 (latest) |
| Processor | Intel 4th gen (Haswell, 2013+) or AMD Zen (Ryzen, 2017+) | Intel 8th gen+ or AMD Ryzen 3000+ |
| CPU feature | AVX2 instruction set | AVX2 (all modern CPUs since 2013) |
| RAM | 4 GB | 8 GB or more |
| Disk space | 500 MB (app + base model) | 2 GB (for larger Whisper models) |
| WebView | WebView2 (pre-installed on Windows 10 1803+) | Automatic |
| Microphone | Any built-in or external mic | USB or built-in mic |
Supported Processors
| Processor | AVX2 Support | Status |
|---|---|---|
| Intel 4th gen+ (Haswell, 2013+) | Yes | Supported |
| Intel 3rd gen or older (2012-) | No | Not Supported |
| AMD Ryzen (all generations) | Yes | Supported |
| AMD FX/A-series (pre-2017) | Varies | Limited |
| Qualcomm Snapdragon X (ARM) | N/A | Not Supported |
How to Check if Your CPU Supports AVX2
- 1Press
Win + R, typemsinfo32, press Enter - 2Look at "Processor" — if it's Intel 4th gen (i3/i5/i7-4xxx) or newer, or AMD Ryzen, you have AVX2
Quick rule: If your PC was manufactured in 2014 or later, it almost certainly supports AVX2.
Windows Model Recommendations
| Model | Performance | Recommendation |
|---|---|---|
| Tiny (75 MB) | Fast | Good for quick use |
| Base (142 MB) | Good | Recommended for daily use |
| Small (466 MB) | Moderate | Good balance |
| Medium (1.5 GB) | Slow | Only on powerful machines (i7-12th gen+, Ryzen 5000+) |
| Large models | Very slow | Not recommended without GPU; consider cloud transcription |
Whisper Models
Disk space and RAM usage by model
| Model | Disk Space | RAM Usage | Best For |
|---|---|---|---|
| Tiny | 75 MB | ~273 MB | Quick commands, drafts, older hardware |
| Base | 142 MB | ~388 MB | Daily use, good balance |
| Small | 466 MB | ~852 MB | Best speed-accuracy balance |
| Medium | 1.5 GB | ~2 GB | High accuracy, powerful hardware |
| Large V3 Turbo | 1.5 GB | ~3 GB | Maximum accuracy, multilingual |
| Large | 2.9 GB | ~4 GB | Research, maximum fidelity |
Cloud Transcription (Optional)
Faster transcription for older hardware
For users who want faster transcription or have older hardware, Vocoding offers optional cloud-based transcription through compatible API providers.
Privacy
Cloud transcription is completely opt-in. By default, all audio processing happens locally on your device.
When to use it
If your hardware is slow with local transcription, or you need maximum accuracy.
Speed
Cloud transcription is typically 5-10x faster than local on most hardware.
Providers & Cost
Configure your own API key in Settings (Groq, OpenRouter, etc.). Groq offers transcription from $0.02/hour.
Not Supported
Configurations that are not currently compatible
| Configuration | Reason |
|---|---|
| macOS 12 (Monterey) or older | Minimum requirement is macOS 13 |
| Macs older than 2017 | Cannot run macOS 13 |
| Windows 32-bit | Not supported |
| Windows ARM (Snapdragon) | Not yet tested; future support planned |
| Linux | Not yet available; planned for a future release |
| iOS / Android | Desktop-only application |
| Chromebook | Not supported |
Frequently Asked Questions
Common questions about compatibility
Does Vocoding work on Intel Macs?
Yes. Intel Macs from 2017 to 2020 running macOS 13+ are supported. Transcription runs on CPU only (no GPU acceleration), so we recommend using the Tiny or Base whisper model for the best experience.
Do I need an internet connection?
No. Vocoding works 100% offline for voice transcription. An internet connection is only needed to download the app and Whisper models initially, and optionally for cloud LLM optimization or cloud transcription.
What if my computer is too slow for local transcription?
You can enable cloud transcription in Settings. This sends your audio to a cloud API for faster processing. Cloud transcription is completely opt-in and disabled by default.
Does it work with external microphones?
Yes. Vocoding works with any microphone your system recognizes — built-in, USB, Bluetooth, or audio interface.
How much disk space do I need?
The app itself is under 100 MB. You'll also need space for the Whisper model you choose: Tiny (75 MB), Base (142 MB), Small (466 MB), Medium (1.5 GB), or Large (2.9 GB).
How much RAM do I need?
4 GB minimum. The Whisper model loads into memory during transcription. The Tiny model uses ~273 MB, Base uses ~388 MB, and Small uses ~852 MB. We recommend 8 GB+ for comfortable use with the Small model or larger.
What about Windows 7 or 8?
Not officially supported. While the app might install, we cannot guarantee stability or performance on Windows versions older than 10.
Will you support Linux?
Linux support is planned for a future release. The app architecture supports it, but we need to complete testing and polish the experience first.
Last updated: February 12, 2026
Ready to get started?
Your system meets the requirements? Download Vocoding and transform your workflow with voice.