CereVault Serve
Local AI Made Easy
Only for Mac
Free
Mac
Your Mac as a private AI server for every device. New: optional web search in the browser chat, drag-and-drop images, and tool-call fixes.
CereVault Serve lets you run private AI on your Mac.
Pick a model from a curated catalog and your Mac becomes a personal AI for your other devices. Your iPhone, iPad, and other Macs on the same Wi-Fi can chat through CereVault Serve over your local network. Nothing leaves your devices unless you ask it to.
Built on Apple's MLX framework, CereVault Serve runs open-weight language models entirely on your Mac. No accounts. No subscriptions. Nothing is sent anywhere unless you turn on the optional web search switch.
KEY FEATURES
• Curated model catalog. Pick from Llama 3.2, Gemma 3 and 4, Qwen3.5, and Phi 4 Mini with one click, including vision models that understand images. Each model downloads directly to your Mac.
• Bring your own model. Add any MLX model from Hugging Face; CereVault Serve checks it will run before downloading.
• Lightweight menu bar app. Lives quietly in your menu bar, with no Dock icon and no windows in your way until you need them.
• Private by default. Models run entirely on your Mac. Network access is loopback-only until you explicitly turn on device sharing.
• Local network sharing. Turn on one toggle to let your iPhone, iPad, and other Macs on your Wi-Fi connect.
• OpenAI-compatible API. Chat, streaming, vision, tool calling, and optional embeddings. Works with any client that supports a custom base URL.
• Easy to find. While sharing is on, your Mac announces itself on your network (Bonjour) so compatible apps can discover it.
• Optional access code. Require a code before anyone on your network can connect.
• Browser chat. Anyone you share with can chat from a web page, and conversations stay on their own device. Turn on optional web search and the page gets a Web switch so the model can look things up.
• Streaming responses. Token-by-token output for natural typing-style conversations.
• Storage you control. Manage downloaded models in one place. Free disk space anytime.
PRIVACY
CereVault Serve runs entirely on your Mac. No analytics. No telemetry. No accounts. Nothing is ever sent to the developer. The only other connections are ones you choose: model downloads from Hugging Face, and the optional web search switch (off by default), which sends only the search terms the model writes to DuckDuckGo.
When you choose to download a model (or turn on the optional Embeddings feature), the files come directly from Hugging Face, a public model registry. That connection is between your Mac and Hugging Face only.
REQUIREMENTS
• Apple Silicon Mac (M1 or later)
• macOS 14 or later
• 1–6 GB free disk space per downloaded model
Ratings & Reviews
- This app hasn’t received enough ratings or reviews to display an overview.
Optional web search for the browser chat, plus fixes.
• Web search in the browser chat (off by default): turn it on in Settings and the chat page gets a Web switch. When it's on, only the search terms the model writes are sent to DuckDuckGo from your Mac, and answers list the sources they used.
• Drag and drop images onto the chat page to attach them (vision models).
• Tool calls from some Gemma models are now recognized correctly, and hidden reasoning text no longer appears in replies.
The developer, Brett Bakker, indicated that the app’s privacy practices may include handling of data as described below. For more information, see the developer’s privacy policy .
Data Not Collected
The developer does not collect any data from this app.
Accessibility
The developer has not yet indicated which accessibility features this app supports. Learn More
Information
- Seller
- Brett Bakker
- Size
- 13.2 MB
- Category
- Productivity
- Compatibility
Requires macOS 14.0 or later.
- Mac
Requires macOS 14.0 or later.
- Mac
- Languages
- English
- Age Rating
4+
- 4+
- Copyright
- © 2026 Brett Bakker