jekalmin/extended_openai_conversation

Home Assistant custom component of conversation agent. It uses OpenAI to control your devices.

What it solves

This project provides an enhanced version of the OpenAI Conversation integration for Home Assistant, allowing an AI assistant to actually perform actions within the smart home environment rather than just chatting. It bridges the gap between a Large Language Model (LLM) and the physical control of devices, automations, and external data sources.

How it works

The integration uses OpenAI's function calling capabilities to map natural language requests to specific Home Assistant services. By exposing entities to the assistant, the LLM can identify which devices are available and call the appropriate services to control them. It supports several function types:

  • Native: Built-in functions for executing services, adding automations, and retrieving state history.
  • Script: Executes a sequence of Home Assistant services.
  • Template: Returns a value based on a template.
  • REST: Fetches data from external API endpoints.
  • Scrape: Extracts information from websites.
  • Composite: Executes a sequence of functions.

Who it’s for

Home Assistant users who want a more powerful, agentic AI voice or text assistant that can manage their home, create automations on the fly, and interact with external web data.

Highlights

  • Direct Home Control: Ability to call Home Assistant services to control devices.
  • Automation Creation: Can generate and add new automations to the system via natural language.
  • External Data Integration: Can retrieve information from external APIs or web pages via REST and scraping.
  • State History: Ability to retrieve and analyze historical data of entities.
  • User Context: Option to pass the current user's name to the AI for personalized interactions.

Related

  • Project
    zhnt/loushang

    A method-native AI work system that provides a CLI workbench and SDK to make complex software development tasks reliable, recoverable, and verifiable.

  • Project
    acon96/home-llm

    A local LLM integration and set of fine-tuned small models that enable private, offline voice and chat control for Home Assistant smart homes.

  • Project
    tigerless-labs/agent-memory

    A long-term memory runtime for AI agents that uses Markdown files as the source of truth and a local index for ranked retrieval, enabling persistence across sessions.

  • Project
    aeon-toolkit/aeon

    aeon is a Python library (scikit‑learn compatible) that provides a unified API for a wide range of time‑series machine‑learning tasks, including state‑of‑the‑art classification, regression, clustering, forecasting, anomaly detection, segmentation, similarity search, transformations, distances/kernels and benchmarking. It bundles both classic statistical models and deep‑learning architectures, is maintained by original algorithm authors, and offers extensive documentation, examples, and community support.