A private AI workspace for your Mac

Run AI models locally on your own machine — offline, with nothing leaving your device. Or bring your own key for cloud models. No account, no tracking, no markup.

Intel & Apple Silicon · macOS 13+

Works with the models you already use

Ollama Claude OpenAI Gemini Grok DeepSeek Qwen Cerebras OpenRouter

+ Groq, LM Studio, MLX, and any OpenAI-compatible API

Private because it runs on your Mac

Most AI apps route everything through someone else's servers. Elvean is a native Mac app — run models locally for full privacy, or connect your own cloud keys when you want more power. You decide where each conversation goes.

Run models locally, fully offline

Download a model inside Elvean and run it on your own hardware — no internet, no API key, no account. You can also connect Ollama, LM Studio, or MLX if you already use them. When you chat with a local model, the conversation never leaves your Mac.

  • Built-in local models you download in the app
  • Also works with Ollama, LM Studio, and MLX
  • Metal-accelerated inference on Apple Silicon
  • Works with no internet connection
Private AI for Mac running local models offline

Bring your own key for cloud models

When a task needs a frontier model, connect your own API key for Claude, GPT, Gemini, Grok, or 300+ models through OpenRouter. You pay the provider directly at their rates — Elvean adds no markup and isn't in the middle. Requests go straight from your Mac to the provider.

  • Bring your own key — pay providers directly, no markup
  • Claude, OpenAI, Gemini, Grok, Groq, OpenRouter, and more
  • Keys stored in the macOS Keychain, never on a server
  • Switch between local and cloud mid-conversation with @mentions
Connecting your own API key for cloud models in Elvean

AI that takes action

Go beyond question and answer. Elvean's agentic tools let models search the web, chain multiple steps together, and connect to external services through MCP. Ask a question, and watch it research, analyze, and present findings.

  • Multi-step agentic tool calling with live progress
  • Web search with inline citations and source links
  • MCP server connections for files, databases, and APIs
  • Image and video search built in
Portfolio breakdown with holdings and key moves

Conversations that stay organized

Long conversations get messy. Elvean brings Slack-style threads to AI — branch off into a side discussion without losing your main context. Fork any message to explore a different direction. Your thinking stays structured.

  • Threaded conversations — branch without losing context
  • Fork any message to explore alternate directions
  • Folders, favorites, and full-text search across all chats
  • Cmd+F find within any conversation
Threaded conversations with stock analysis

A native app, not a browser in disguise

Elvean is built in SwiftUI — not a web app wrapped in Electron. It launches in under a second, stays light on memory, and behaves like a Mac app should. Local inference runs on Metal, and the whole thing works on both Apple Silicon and Intel.

  • Native SwiftUI — Apple Silicon and Intel, macOS 13+
  • System-wide prompt panel: select text anywhere, hit a hotkey, get AI
  • Prompt library with variable placeholders and Cmd+/ access
  • Per-conversation settings: temperature, tokens, system prompts
Travel itinerary with rich content and embedded photos

What "private" actually means here

No marketing hand-waving — here's exactly how your data is handled. Local models run on your Mac and never touch the network. Cloud requests go straight from your Mac to the provider you chose, never through us. Conversations are stored on your device; API keys live in the macOS Keychain. Elvean has no analytics or telemetry, and you never make an account.

And there's more

🎙Voice input & text-to-speech
📎Image & PDF attachments
📝Edit any AI response in Markdown
📊Token usage tracking
📤Export to Markdown, HTML, PDF
Metal-accelerated local inference

Free to start, Pro when you're ready

No account required. Download and start chatting immediately.

Free

$0

forever

  • 300+ cloud models (BYOK)
  • Ollama, LM Studio & MLX local models
  • Unlimited conversations
  • Folders & organization
  • System-wide prompt panel
  • Fork any message
  • Prompt library
Download Free
Most Popular

Pro

$4.99/mo

or $49.99/year (save 17%) · 7-day free trial on all plans

  • Everything in Free, plus:
  • Web, image & video search
  • Interactive charts, diagrams & maps
  • Screenshot to AI
  • Apple Calendar & Weather tools
  • Threads & @mention model switching
  • Markdown editing of responses
  • Unlimited MCP servers
  • Export (Markdown, HTML, PDF)
  • Priority support
Start Free Trial

Common questions

What's the most private AI app for Mac?

Elvean runs models locally on your Mac, so conversations can stay entirely on your device with no internet connection. It collects no telemetry or analytics and needs no account. That combination — local models, no tracking, no sign-up — is what makes it a genuinely private option.

Can I use AI on my Mac completely offline?

Yes. Download a model in the app (or connect Ollama, LM Studio, or MLX) and you can chat with no internet at all. Nothing is sent anywhere — the model runs on your own hardware.

Do I have to create an account?

No. There's no account, no sign-up, and no email. Download it, open it, start chatting.

What happens when I use a cloud model?

The request goes directly from your Mac to the provider you picked (OpenAI, Anthropic, Google, etc.) using your own API key. It doesn't pass through Elvean's servers, and there's no markup — you pay the provider their rate. Your key stays in the macOS Keychain.

Is it really free?

Yes. Local models and bring-your-own-key cloud models work on the free tier with no time limit. Pro ($4.99/mo) adds features like web search, interactive charts and maps, and MCP servers.

Your AI, on your Mac, on your terms

Download, open, chat. No account, no tracking, no sign-up.

Download Free for Mac Intel & Apple Silicon · macOS 13+ · Free