Agent loaded but a local LLM is required to respond.
1. Download: ollama.com
2. Launch with CORS enabled:
OLLAMA_ORIGINS=* ollama serve
3. Pull the model (first run only):
ollama pull llama3.2:3b
>> Reload page once Ollama is running.