Native on iPhone, iPad, Mac, and Vision Pro
Private intelligence, entirely yours.
Run advanced models, search your own files, and use real tools without giving up your conversations.
Download on the App StoreThe privacy model
Local by default.
Local chats can remain on your device.
Optional network routes are visibly controlled.
Off-grid Mode blocks external HTTP and HTTPS.
One private system, wherever you think.
Current product surfaces. Real local models.
Load local models, Apple Foundation Models, and support models from one native library.
A complete local AI workspace.
Run local models
GGUF, MLX, Core ML, ExecuTorch, and Apple Foundation Models in one native client.
02Work from your files
Read PDFs more thoroughly and ground answers in your own datasets with retrieval that stays on device.
03Use real tools
Open web evidence, inspect citations, and connect MCP tools when the task needs them.
04Connect your stack
Use LM Studio, Ollama, OpenAI-compatible endpoints, or Noema Relay without changing clients.
Noema's first model
Small enough for the edge. Built to reason with less.
Noema 2B is a 2-billion-parameter model tuned for token-efficient reasoning, mathematics, and clear instruction following on personal hardware.
Explore Noema 2B- Parameters
- 2 billion
- Base
- Qwen3.5-2B
- Released
- July 2026
Your models. Your files. Your device.
No subscription. No cloud dependency.
Get Noema