Feature request: Add Ollama as a local model provider in QuickStart

Context

I’m trying to set up OpenClaw locally on a Linux notebook.

The current QuickStart flow only allows selecting remote providers (OpenAI, Anthropic, etc.), which blocks a local-first setup.

Problem

• No option to select a local model provider
• No Ollama integration in the onboarding wizard
• Difficult or unclear local setup on Linux

As a result, I haven’t been able to complete a fully local installation using the wizard.

Requested feature

Add Ollama as a first-class model provider in the QuickStart flow.

Suggested behavior

• Allow selecting ollama as a provider during onboarding
• Automatically detect a running Ollama instance at localhost:11434 (if available)
• Provide basic local model selection (e.g. llama3, mistral, codellama)

Expected benefit

• Improved Linux support
• Fully local, offline-capable setup
• Better alignment with a local-first workflow

Thanks for the great project 🚀