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

# Model Download & Management

> Download, manage, and configure transcription models.

## Your First Model

Vowen offers you a transcription model to download on first launch. The **Base** model is recommended by default and works well for everyday use across English and multilingual contexts. You can switch to a different model anytime from **Settings > Models**.

## Downloading Additional Models

<Steps>
  <Step title="Open Models Section">
    Go to **Settings > Models** in the Vowen settings window.
  </Step>

  <Step title="Browse Available Models">
    You'll see all available models with their size, language support, and speed ratings.
  </Step>

  <Step title="Download">
    Click the download button next to any model. Progress is shown as a percentage with download speed.
  </Step>

  <Step title="Start Using">
    Once downloaded, the model is automatically selected. You can switch between downloaded models at any time.
  </Step>
</Steps>

## Managing Models

### Switching Models

Click on any downloaded model to make it your active transcription model. You can also switch models quickly from the system tray right-click menu.

### Deleting Models

To free up disk space, click the delete icon next to any downloaded model.

### Disk Space

| Model              | Disk Space Required                 |
| ------------------ | ----------------------------------- |
| Tiny / Tiny.en     | \~78 MB                             |
| Base / Base.en     | \~148 MB                            |
| Small / Small.en   | \~488 MB                            |
| Medium / Medium.en | \~1.5 GB                            |
| Large v3           | \~3 GB                              |
| Large v3 Turbo     | \~1.6 GB                            |
| Parakeet TDT 0.6B  | \~1 GB (macOS) / \~478 MB (Windows) |

## GPU Acceleration (Windows)

If you have an NVIDIA GPU:

1. Scroll to the bottom of the Models page
2. Find the "GPU Acceleration" card
3. Click **Download** to get the CUDA acceleration modules
4. Restart Vowen after download completes

<Note>GPU acceleration is currently supported for NVIDIA GPUs only. AMD and Intel GPUs are not yet supported.</Note>

## Resource Efficient Mode

If you're on a resource-constrained machine, enable **Resource Efficient Mode** in Settings > General (at the bottom). This uses the CLI method instead of keeping a persistent server in memory: lower RAM usage with slightly slower transcription.

## Cloud Models

Cloud models do not need to be downloaded. To use one, sign up with the provider, generate an API key, and paste it into Vowen.

<Steps>
  <Step title="Pick a provider">
    Open **Settings > Models** and choose a cloud model from the list.
  </Step>

  <Step title="Sign up and get an API key">
    Click through to the provider's signup page. Most cloud transcription providers (Groq, Deepgram, AssemblyAI, Gemini, and more) offer a generous free tier; a payment method is only required if you exceed it.
  </Step>

  <Step title="Connect">
    Paste the API key into Vowen. The model is ready to use immediately.
  </Step>
</Steps>

Cloud models consume no local disk space. You can connect as many providers as you want and switch between them at any time.

***

Have a question about downloading or connecting models? See the [Models FAQ](/faq#models).