AI Automation Suggester

AI Automation Suggester helps Home Assistant users find automation ideas based on their home's devices, rooms and existing automations. It produces YAML suggestions for review rather than installing automations itself.

• Suggestions based on your home: The integration uses device capabilities, states and area information to propose automations relevant to your setup.
• Goals and filters: Custom prompts can focus ideas on goals such as presence lighting, while filters control which entities the model considers.
• Reviewable YAML: It checks returned YAML and records warnings for missing or potentially truncated output so you can assess suggestions before using them.
• Suggestion history: Notifications and dashboard cards display ideas, while stored history keeps generated YAML, provider and model details, warnings and review status.

The integration runs in your Home Assistant instance under the MIT open-source license. Generation uses a separate AI backend. Ollama and LocalAI support local model processing; cloud providers such as OpenAI, Anthropic, Google and Groq receive the household context included in generation requests. It also supports LiteLLM and OpenAI-compatible endpoints.

You can mark suggestions as accepted, declined or dismissed. Generation can run on demand, but scheduled reviews and reactions to newly added devices require a Home Assistant automation. The integration doesn't create those automations or a schedule.