YOUR WORKSPACE
AI assistants
Local models, Hermes Agent and OpenClaw in your desktop.
Choose your assistant
Open Settings → AI Assistants → Add assistant. Follow the type, source, model and review steps. Use a local model server or connect Hermes Agent or OpenClaw. Local chat can discover Ollama, LM Studio, llama.cpp and vLLM on the host.
Models and resources
An administrator can install a shared EveryOS Ollama with Docker and download models suited to available memory. NVIDIA or AMD acceleration is offered only when the host and Docker support it. Review GPU permissions and storage requirements before installation.
Control your conversations
Set the model, response language, personality and answer length. Chat histories and keys are encrypted on your server. Direct chat with a local model on the host stays local; a cloud provider or remote model sends requests to that configured service.
Existing agents and chat history
AI access is currently administrator-only. Discover existing agents and import supported conversation history without changing the original files. Favorite or archive chats, change models and use the AI chat widget. Switching a managed OpenClaw model restarts it after confirmation; external OpenClaw configuration is changed in that agent itself.
Based on EveryOS v0.1.38
View source · ai-agents.md