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

# System requirements for installing Whatsy on macOS

> Check what your Mac needs before installing Whatsy: macOS 14 or later, Google Chrome, and a locally-running Ollama instance with at least one model pulled.

Whatsy bundles most of what it needs inside the app, but a few dependencies live outside the bundle because they're either too large to ship or deliberately kept separate for your privacy. Before you install Whatsy, check that your Mac meets all of the requirements below — the agent won't start if any of them are missing.

## System requirements

| Requirement       | Minimum version     | Notes                                                                      |
| ----------------- | ------------------- | -------------------------------------------------------------------------- |
| **macOS**         | 14 (Sonoma)         | Required for the native app to run                                         |
| **Google Chrome** | Any current release | Required by the WhatsApp web bridge; see note below                        |
| **Ollama**        | Latest              | Must be running at `http://localhost:11434` with at least one model pulled |

## Google Chrome

Whatsy connects to WhatsApp through a browser-based session that runs silently in the background. Chrome handles this session invisibly — you won't see a browser window open.

Whatsy looks for Chrome at its default macOS path:

```text theme={null}
/Applications/Google Chrome.app/Contents/MacOS/Google Chrome
```

<Warning>
  If Google Chrome is not installed, Whatsy cannot link to WhatsApp. The app will show an inline prompt when it detects Chrome is missing. Install Chrome from [google.com/chrome](https://www.google.com/chrome) and restart the app before continuing.
</Warning>

## Ollama and model setup

Whatsy generates every reply using a language model that runs entirely on your Mac through [Ollama](https://ollama.com). No request is sent to any cloud inference service.

Ollama must be running and reachable at `http://localhost:11434` before you start the agent. You also need at least one model pulled and ready.

### Recommended models

| Model            | Size   | Why we recommend it                                                        |
| ---------------- | ------ | -------------------------------------------------------------------------- |
| **`llama3.2`** ✅ | \~2 GB | Best balance of tone quality, response speed, and memory use for most Macs |
| `llama3.1`       | \~4 GB | Slightly more capable; better for complex persona instructions             |
| `mistral`        | \~4 GB | Good alternative if you prefer Mistral's output style                      |

We recommend **`llama3.2`** for most users. The model you choose directly affects how well the agent captures your tone — a larger, more capable model generally produces more natural-sounding replies, but it also requires more RAM and will be slower on older hardware. Experiment to find the right balance for your machine.

To pull the recommended model, run:

```bash theme={null}
ollama pull llama3.2
```

To start the Ollama server (required every time your Mac restarts unless you configure Ollama to launch at login):

```bash theme={null}
ollama serve
```

For help installing Ollama or troubleshooting connection issues, see [Ollama setup](/troubleshooting/ollama-setup).

## Pre-flight checklist

Work through these checks before opening Whatsy for the first time:

<Steps>
  <Step title="Confirm your macOS version">
    Go to **Apple menu → About This Mac**. Your macOS version must be **14.0 or later**. If it's older, use **Software Update** to upgrade before continuing.
  </Step>

  <Step title="Install Google Chrome">
    Open Finder and look for **Google Chrome** in your Applications folder. If it's not there, download it from [google.com/chrome](https://www.google.com/chrome) and drag it to Applications.
  </Step>

  <Step title="Install and start Ollama">
    Download Ollama from [ollama.com](https://ollama.com) and install it. Then open Terminal and run:

    ```bash theme={null}
    ollama serve
    ```

    Leave this running in the background.
  </Step>

  <Step title="Pull a model">
    In a new Terminal tab, pull the recommended model:

    ```bash theme={null}
    ollama pull llama3.2
    ```

    Wait for the download to complete before opening Whatsy.
  </Step>

  <Step title="Open Whatsy">
    Launch Whatsy from your Applications folder. The onboarding flow checks for Ollama and Chrome automatically — if anything is missing, it tells you what to fix before you can proceed.
  </Step>
</Steps>

## Storage location

All Whatsy data — your persona, behavior rules, WhatsApp session files, and agent logs — is stored at:

```text theme={null}
~/Library/Application Support/Whatsy/
```

Nothing in this directory is ever uploaded or shared. You can back it up, inspect it, or delete it at any time.