> ## Documentation Index
> Fetch the complete documentation index at: https://elatoai.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Gemma

> Use Google Gemma from mlx-community as the local LLM for your ESP32 AI toy.

Gemma 4 models (26B A4B and 12B in the README) from [`mlx-community`](https://huggingface.co/mlx-community) work with OpenToys. Open the **Models** page, paste the Hugging Face repo name of the Gemma build you want and OpenToys loads it with `mlx-lm` (or `mlx-vlm` if the model needs it).

> [!TIP]
>
> Pick a 4-bit build to keep responses fast.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.