ai-agents-llm-connection-querying-with-ollama-llm-fails-with-httpcore-conne-fcd3ffa1

THEN Ensure the Ollama server is running (`ollama serve`) and the required model (e.g., mistral) is pulled (`ollama pull mistral`). Also check that the Ollama service is accessible on the expected host/port; if running remotely, set the base URL via `Ollama(base_url='...')`. Implement retry logic for transient network issues.

import subprocess
try:
    subprocess.run(['ollama', 'list'], check=True, capture_output=True)
    print('Ollama server is running.')
except subprocess.CalledProcessError:
    print('Ollama server not running. Start with: ollama serve')
    exit(1)