cc-ollama

Install local AI Ollama (Claude Code version)

1. Download and install:https://ollama.com/download
2. ollama --version
3. Pull model:ollama pull llama3.2:latest or ollama pull gemma3:latest
4. If GPU acceleration confirmation is needed:ollama run llama3.2:latest "hello"

Optional: Add Claude Code MCP

If you want Claude to call the local model:

"mcpServers": { "ollama": { "command": "npx", "args": ["-y", "@ollama/mcp-server"] } }

Report: Ollama version, model list, GPU status.