macOS 11+ · Intel & Apple Silicon

A private AI assistant
that lives on your Mac.

Agentio runs open-source LLMs locally through Ollama. Your chats, documents and prompts never leave the device. No accounts, no telemetry, no cloud.

Download for macOS See it in action

Free · Open source · ~2 MB installer

Runs the models you already trust

Llama 3 Mistral Phi-3 Gemma 2 TinyLlama Qwen 2.5

Everything you need. Nothing you don't.

A clean, native macOS experience built around your privacy.

Truly private

Every prompt and reply stays on your Mac. No network calls to the cloud, ever.

Real chat, real models

Streaming replies from local LLMs via Ollama. Switch models mid-conversation.

Markdown & code

Rich rendering with syntax highlighting for Swift, Python, JS, Go, Rust and more.

Full control

Tune temperature, top-p, top-k, repeat penalty and system prompts per model.

Light & Dark

Beautiful themes, custom accent colors, native macOS look and feel.

Telegram bridge

Chat with your local AI from any device through a Telegram bot.

Take a look

Agentio is designed to feel like a native Apple app.

Chat with a real reply from a local model
Chat with real local models. Streaming replies, markdown, code.
Onboarding with confetti
Delightful onboarding with confetti animation.
AI models shop with search
Models shop with live search and progress.
Settings in dark mode
Deep settings in a beautiful dark theme.
Developer console
Developer mode with live logs.
DMG installer
Drag-and-drop installer with beautiful artwork.

Install in 3 minutes

Zero accounts. Zero telemetry. Just your Mac and your data.

1

Install Ollama

Agentio uses Ollama as its local model runtime.

brew install ollama
2

Download Agentio

Grab the DMG, drag the app to Applications.

Download .dmg
3

Start chatting

Pull a model and start talking. Agentio auto-detects the runtime.

ollama pull tinyllama

Ready when you are.

Download Agentio

macOS 11 Big Sur or later · Intel and Apple Silicon