Search intent answer
This is a high-friction failure because the avatar may speak the error repeatedly while the useful detail sits in backend logs. The diagnostic maps the symptom to the likely LLM connection cause.
How the console handles it
- Paste the backend log or the spoken frontend error.
- Mark whether you use Ollama, LM Studio, OpenAI-compatible API, Claude, Gemini, or Groq.
- Check base URL, API key presence, model name, and rate limit signals.
- Generate a ranked fix list.
- Unlock a full troubleshooting receipt for your setup.