Guides

Local AI setup guides

Narrow, sourced answers to the setup failures people hit after the fit verdict: a model that will not load, a context setting that costs more memory than expected, or a vendor feature whose requirements are easy to misread.

Local AI hardware buying guide →

Choose an upgrade, a complete system, or a rental before spending.

Updated 2026-10-19

Ollama Out of Memory? Read `ollama ps` First

Ollama OOM or a model that loads but crawls? Read the `ollama ps` processor/context split, see why context length changes memory, and choose the smallest fix.

Updated 2026-10-19

Qwen3.8-27B + Splash on Apple Silicon: Requirements & Setup

Splash in LM Studio Bionic requires an M3+ Mac, macOS 26.4+, 36GB unified memory (48GB+ recommended), and Bionic 1.1.5+. Setup steps and TokenAssemble's no-verdict boundary.

Need a verdict for your exact machine? Run the checker — fit, grade, context, and runtime gates, with every assumption shown.