Device-first guides
Start with Android, iPhone and iPad, Mac, or Windows. The goal is to match the docs to the hardware you actually have.
Gemma 4 offline guides
Pick the right model, choose the right runtime, and get Gemma 4 running locally on the hardware you already own.
Runtime selection
Compare AI Edge Gallery, LM Studio, llama.cpp, and Ollama without reading four separate docs sets first.
Troubleshooting
Fix out-of-memory errors, slow loading, and missing GPU or NPU acceleration fast.
Model fit
Learn when to use E2B, E4B, 26B A4B, or 31B instead of defaulting to the biggest download.
Open the quick start The shortest path from “I just found Gemma 4” to a working local install.