Your AI Assistant, Running Locally
Adaptive Layering Intelligence System. Native on iOS & macOS with local LLMs, Ollama endpoints, and Apple Intelligence. Choose separate brains for chat and automations. Your data stays yours.
Built for the Real World
Six pillars that make ALIS a complete, trustworthy AI assistant you can rely on every day.
Flexible AI Engine
Local LLMs via llama.cpp, external Ollama endpoints, or Apple Intelligence on supported devices. Pick a different brain for chat and for automations.
Modular Skills
Gmail, Telegram, Calendar, Apple Mail — each skill is a self-contained actor with its own manifest, tools, and lifecycle.
Event-Driven Automations
Trigger actions on email received, schedule, or manual input. Multi-action sequential flows with consent control.
Runtime Truth
ALIS knows its own state: which skills are active, authenticated, and healthy. No guessing, no hallucination.
Native Voice
Apple Speech-to-Text and Text-to-Speech for a complete voice pipeline. Talk to ALIS, ALIS talks back.
Live Data
Weather, news, markets, sports — temporal facts from real sources. The model reasons; connectors verify.
Choose Your Model
Download only what you need. Each model runs entirely on your device — no cloud, no subscription, no compromise.
Qwen 0.8B
Lightweight and fast. Perfect for quick answers on any device.
Qwen 2B
The sweet spot. Solid reasoning without breaking a sweat.
Qwen 4B
Code, reasoning, conversation — a real workhorse.
Qwen 9B
Deep analysis and complex tasks. For when depth matters.
Llama 3.1
Meta's open model. Strong across the board.
Apple Intelligence
Already on your device. Zero download, instant start.
Works Instantly
On supported devices, ALIS works the moment you open it. Apple Intelligence is already there — no downloads, no setup, no waiting.
Then download premium models whenever you want. More power, same privacy.
Two Brains. One Assistant.
Choose a lightweight model for everyday chat and a powerful one for automations. Each role gets the right brain.
Chat Brain
The brain that talks to you
Fast, conversational, always ready. Pick a small model for instant responses.
Automation Brain
The brain that acts for you
Precise, methodical, reliable. Pick a powerful model for complex multi-step tasks.
Built Different
How ALIS Works
A layered architecture where the core is a pure orchestrator. Skills self-describe, self-register, and own their entire domain.
Core Principle
The orchestrator knows contracts, not domains. Zero core contamination — every new skill plugs in without touching the brain.
Ready to Meet ALIS?
Join the waitlist to be among the first to experience AI that truly runs on your device. No cloud dependency. No data leaving your hands.
No spam, ever. We'll only email you when ALIS is ready.