> ## Documentation Index
> Fetch the complete documentation index at: https://docs.voicetypr.com/llms.txt
> Use this file to discover all available pages before exploring further.

# System Requirements

> Hardware and software requirements for running VoiceTypr

VoiceTypr runs on macOS and Windows with minimal system requirements. However, certain features require specific configurations.

## macOS Requirements

### Operating System

* **macOS 13.0 (Ventura) or later** is required
* VoiceTypr is fully signed and notarized by Apple
* Runs natively on both Intel and Apple Silicon Macs

### Disk Space

* **3-4 GB of free disk space** for AI models
* Model sizes vary:
  * Tiny models: \~75 MB
  * Base models: \~142 MB
  * Small models: \~466 MB
  * Medium models: \~1.5 GB
  * Large models: \~2.9 GB

### Required Permissions

<AccordionGroup>
  <Accordion title="Microphone Access" icon="microphone" defaultOpen>
    Required to record your voice for transcription.

    **How to grant:**

    1. Open VoiceTypr
    2. Go to Advanced settings
    3. Click "Grant" for Microphone permission
    4. Allow in System Settings when prompted
  </Accordion>

  <Accordion title="Accessibility Access" icon="universal-access">
    Required for global hotkeys to work system-wide and for automatic text insertion.

    **How to grant:**

    1. Open VoiceTypr
    2. Go to Advanced settings
    3. Click "Grant" for Accessibility permission
    4. Add VoiceTypr in System Settings > Privacy & Security > Accessibility
  </Accordion>
</AccordionGroup>

<Note>
  Without Accessibility permission, hotkeys won't work when other apps are in focus.
</Note>

## Windows Requirements

### Operating System

* **Windows 10 (64-bit)** or **Windows 11**
* No special permissions required - works out of the box

### Disk Space

* **3-4 GB of free disk space** for AI models
* Same model sizes as macOS (see above)

### GPU Support

VoiceTypr automatically detects and uses your GPU for **5-10x faster** transcription:

<CardGroup cols={3}>
  <Card title="NVIDIA" icon="n">
    GeForce, RTX, or Quadro series GPUs

    [Update drivers →](https://www.nvidia.com/drivers)
  </Card>

  <Card title="AMD" icon="a">
    Radeon series GPUs

    [Update drivers →](https://www.amd.com/support)
  </Card>

  <Card title="Intel" icon="i">
    Intel Iris Xe or Arc series GPUs

    [Update drivers →](https://www.intel.com/content/www/us/en/support/products/80939/graphics.html)
  </Card>
</CardGroup>

<Info>
  VoiceTypr automatically falls back to CPU if GPU acceleration is unavailable.
</Info>

## Hardware Acceleration

### Metal (macOS)

* Automatic hardware acceleration on all Macs
* Uses **Metal GPU** for Whisper models
* Uses **Apple Neural Engine** for Parakeet models (Apple Silicon only)

### GPU (Windows)

* Automatic detection and usage when available
* **5-10x faster** than CPU-only transcription
* Requires up-to-date graphics drivers

<Warning>
  Keep graphics drivers updated for best performance. Outdated drivers may prevent GPU acceleration.
</Warning>

## Microphone Requirements

Any microphone works with VoiceTypr:

* Built-in laptop microphone
* USB microphone
* Bluetooth headset
* External audio interface

<Tip>
  For best accuracy, use a quality microphone in a quiet environment. Built-in mics work well for most use cases.
</Tip>

## Network Requirements

<Info>
  VoiceTypr works **100% offline** for core transcription features.
</Info>

Internet connection is only required for:

* Initial app download and installation
* Downloading AI models
* AI Enhancement features (Groq/Gemini)
* License activation
* Automatic updates

Once models are downloaded, transcription works completely offline.