Local AI Overview

Yumi supports running AI models locally on your machine for private, offline conversations.

How It Works

Local AI mode connects to a local model server running on your machine. You can use your own models and API keys (BYOK) for fully private AI interactions.

All chats in local mode are stored on your device.

Supported Setups

• Local model servers (llama.cpp, Ollama, LM Studio, any OpenAI compatible server, etc.)
• Bring Your Own Key (BYOK) for third-party APIs

Privacy

When using local chat, all inference happens on your device or local server. If you're using BYOK, then inference will happen on the respective provider. No data is sent to Yumi's servers when using local chat.