What's new
Announcements
Launches, new models, and capabilities for local AI on your Mac.
Be your own LLM provider
Run local models and route any OpenAI-compatible provider — from one endpoint on your machine.
Read →Bonsai 27B lands in the Model Store
A full 27B mind in 3.8 GB — true 1-bit ternary Qwen3.6-27B that runs on mainstream Macs.
Read →Your local models can now use tools
Aether's local engine now speaks MCP — read files, search the web, run commands, all on your Mac. No cloud, no proxy.
Read →