Your AI.
On your terms.
Open models on your iPhone and Macs, brought together in one familiar chat app. Everyday questions, serious writing, and code—powered by hardware you control.
Apple first · Open model choice · No per-request model fees
One Mac. Or Mac Pool.
Lightspeed.
Everyday replies.
Right here.
Feels like chat.
Runs on iPhone.
Write a better message. Untangle an idea. Plan your day. Lightspeed gives you quick replies from a model running on iPhone itself.
Start with iPhone. Add a Mac when you want more.
Make this sound warmer: “Running five minutes late.”
“I’m running about five minutes late. Thank you for waiting.”
Bigger work.
Your own Macs.
Think through code, longer writing, and harder questions with a model running on your Mac. Reach it from iPhone, iPad or browser. Mac Pool, developed by Stratumus, shares a supported model across compatible Macs—so you can run something larger than one Mac can hold alone.
A reasoning model shared across two Macs
16 GB + 24 GB · Combined memory
Your devices generate the answers. Account chats still pass through Stratumus and are stored in your account. See the data boundaries ↗
One model. Multiple Macs.
Share a supported model across compatible Macs on the same home network.
Illustrative view · Pooling adds capacity; speed varies.
Your local AI.
More connected.
Already use oMLX or Ollama? Stratumus works with them on individual Macs and brings the experience around your devices into one place.
Take your AI with you.
Chat with your Mac from iPhone, iPad or browser. Your Mac connects out, so you don’t have to configure your router.
Control it from your phone.
Choose assistants, manage supported model downloads, and see which Macs are available.
One model. Multiple Macs.
Mac Pool is technology developed by Stratumus. It shares a supported model across compatible Apple silicon Macs on the same home network, giving larger models more room.
Clear by design.
Why use Stratumus with oMLX or Ollama?+
oMLX and Ollama are capable local AI tools. If you need a model server or API for your own apps and coding tools, using them directly may be exactly right. Stratumus adds a connected chat experience across your Apple devices, remote Mac and model controls, and Mac Pool. It supports oMLX and Ollama on individual Macs; Mac Pool is developed by Stratumus to share one supported model across compatible Macs.
Does my code stay entirely on my Mac?+
The model processes your code on your Mac when that Mac answers. Stratumus does not fall back to a cloud AI provider to generate the answer. Normal account chats still pass through the Stratumus service, encrypted in transit, and conversation history and opt-in memories are stored in your Stratumus account. Local inference and entirely local data storage are different things. Read our privacy policy.
Can I use it without a Mac?+
Yes. The iPhone app can answer with an on-device assistant, including downloaded models such as Lightspeed. A Mac adds access to models that need more capacity.
What does Mac Pool add?+
Mac Pool is technology developed by Stratumus. It splits one supported model across compatible Apple silicon Macs on the same home network, expanding which models you can run. A 32-billion-parameter reasoning model has run across 16 GB and 24 GB Macs using Mac Pool. Pooling does not guarantee faster replies; speed depends on the model, hardware and connection. Mac setup currently requires Ollama or oMLX.
What does “open models” mean?+
Models you can download and run on your own hardware. You choose a supported model; generating its replies has no per-request model fee and needs no paid cloud AI API key. Each model has its own license, which may include conditions for commercial use. “Open models” describes the models; Stratumus’s own source code is not currently public.
Make AI
your own.
For everyday questions. For the work you’re building.
Join early access and we’ll tell you when it’s ready.