gritaro/gigachain

This custom component for Home Assistant allows you to generate text responses using GigaChain LLM framework (like GigaChat, YandexGPT, ChatGPT, Anyscale).

🦜️🔗 GigaChain (GigaChat + LangChain)

GigaChain integration with Home Assistant

This integration implements Voice Assistant for Home Assistant using GigaChain framework. Currently supported LMMs:

• GigaChat (Sber LLM)
• YandexGPT
• OpenAI aka ChatGPT (not tested)
• Anyscale

Installation

Install it like any other HACS integration.

Requirements

Home Assistant with installed HACS

Installation with HACS

Find GigaChain in HACS store. If you can't find it in store, you could add this url as HACS custom repository.

Restart Home Assistant.

Add Integration

After adding, configure integration.

Settings

GigaChat

GigaChat Authorization

You need to register at https://developers.sber.ru/studio and get an "authorization data" key.

Note

You can find more details in GigaChat official documentation.

YandexGPT

Quick start

Create service account with role ai.languageModels.user. Create API key. You can find Folder ID using this link.

OpenAI

Create API key here https://platform.openai.com/account/api-keys

Anyscale

Register account and create API key here

Configuration

• Prompt template (template, Home Assistant template)

The starting text for the AI language model to generate new text from. This text can include information about your Home Assistant instance, devices, and areas and is written using Home Assistant Templating. Default value comes from official integration OpenAI Conversation

• Model (model, string)

Language model is used for text generation

• Temperature (temperature, float)

A value that determines the level of creativity and risk-taking the model should use when generating text. A higher temperature means the model is more likely to generate unexpected results, while a lower temperature results in more deterministic results.

• Max Tokens (max_tokens, int)

The maximum number of words or “tokens” that the AI model should generate in its completion of the prompt.

• Process HA Builtin Sentences (process_builtin_sentences, bool)

If enabled, integration first will pass all sentences to HA built-in sentence processor. This is default behaviour of default Home Assistant Voice Assistant engine which allow you to use commands something like turn on the living room light. If sentence will not be recognized by HA, it will be passed further to chosen LLM.

• Chat History (chat_history, bool)

Keep all conversation history.

Using as Voice Assistant

Create and configure Voice Assistant: