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

# macOS First Run

> What the HyperWhisper first-run setup asks for on macOS, and what Set Up Later keeps.

The first time that you open HyperWhisper on a Mac, the app shows a setup flow in eight steps. The flow gets the permissions, selects where your speech becomes text, and makes one test recording.

The flow opens on a new installation only, and one time. If you upgrade from an older build that already has modes, HyperWhisper marks the setup as complete and does not show the flow.

## The eight steps

<Steps>
  <Step title="Welcome">
    A summary of the three tasks: select how transcription runs, check your microphone, and make one recording. Click **Get started**.
  </Step>

  <Step title="Two quick permissions">
    * **Microphone** — required. You cannot continue without it. HyperWhisper captures audio only while a recording runs.
    * **Accessibility** — optional. With this permission, HyperWhisper puts the transcript at your cursor. Without it, the transcript goes to the clipboard.

    If you refused a permission before, macOS does not ask again. The step gives an **Open Settings** button instead. HyperWhisper examines both permissions again when you come back to the app.
  </Step>

  <Step title="How do you want to transcribe?">
    Select one of three sources:

    | Source                 | What it uses                                                   |
    | ---------------------- | -------------------------------------------------------------- |
    | **HyperWhisper Cloud** | Your prepaid [credits](/docs/cloud-credits). No API key of your own |
    | **On this Mac**        | A local model. No network after the download                   |
    | **Your API Key**       | Your own account with a cloud provider                         |

    To change the source later, you use the [mode editor](/docs/transcription-modes). You do not do this setup again.
  </Step>

  <Step title="Configure the source">
    The step changes with the source that you selected:

    * **HyperWhisper Cloud** — paste your access key. Click **Test access key**. The card then shows the credits of that key.
    * **On this Mac** — select one of four models: **Parakeet V2** (English only, fastest), **Parakeet V3** (multilingual), **Whisper Base** (smallest download), or **Whisper Large v3 Turbo** (most accurate).
    * **Your API Key** — select a provider and paste your API key. Click **Test API key**. macOS holds the key in your login keychain.

    <Note>
      For **Your API Key**, you must click **Test API key**, even when the key is already in your keychain. HyperWhisper does not accept a key that it did not verify in this session.
    </Note>
  </Step>

  <Step title="Set up the source">
    The step completes the work: it activates the Cloud key, downloads the local model, or writes the API key. It shows the result as a list of checks. A download failure or an activation failure appears here as a message.
  </Step>

  <Step title="Say something. Watch the bars.">
    Select your input device and speak. The level bars move when HyperWhisper hears you. **System Default** is the first row, and it follows macOS, including a headset that you connect later. The **Sound Settings** button opens the macOS sound settings.
  </Step>

  <Step title="Make one recording.">
    Press your recording shortcut one time to start and one time to stop. As an alternative, click the **Record** button. The step gives a sample line to say. The transcript appears below it.

    <Note>
      HyperWhisper does not paste this test transcript into other applications. The text stays in the setup window.
    </Note>
  </Step>

  <Step title="You are ready to dictate.">
    A summary of your transcription source, your microphone, your text delivery (cursor or clipboard), and your shortcut. Click **Done Onboarding**.
  </Step>
</Steps>

## Set Up Later

Each step, except the last one, has a **Set Up Later** button. The button closes the setup flow and puts back the state that the flow changed:

* the default mode and the mode selection
* any provider API key that a test in this flow replaced
* your microphone selection

Two things stay:

* **Models that you downloaded.** The files stay on disk. Use the [Model Library](/docs/models) to remove them.
* **A HyperWhisper Cloud key that you activated.** Activation is an action on your account, not a local setting.

<Warning>
  **Set Up Later** also marks the setup as complete. The flow does not open again. To configure HyperWhisper after this, use **Settings** and the [mode editor](/docs/transcription-modes).
</Warning>

## Next steps

<CardGroup cols={2}>
  <Card title="Permissions" href="/docs/permissions">
    Grant Accessibility later, and read what each permission does.
  </Card>

  <Card title="Keyboard Shortcuts" href="/docs/keyboard-shortcuts">
    Change the shortcut that starts and stops a recording.
  </Card>

  <Card title="Transcription Modes" href="/docs/transcription-modes">
    Change your provider, your model, and your post-processing.
  </Card>

  <Card title="Models" href="/docs/models">
    Compare the local models and the cloud models.
  </Card>
</CardGroup>