Run AI Locally. Zero API Costs.
Anthropic raised prices. You don't have to pay them.
Set up Gemma 4 (26B parameters) in LM Studio and wire it to OpenClaw or Hermes in under 5 minutes — a complete local AI stack that runs entirely on your machine.
What's Inside
Everything in the swipe file
LM Studio Install
Get the local model runner set up on your Mac in under 2 minutes.
Gemma 4 26B Download
17.99GB, 26B parameters, vision-capable — pulled directly in-app.
Local Server Config
OpenAI-compatible API running at 127.0.0.1:1234 — no cloud, no latency.
Hermes Agent Install
One-line curl install. Works on macOS, Linux, and WSL2. No prereqs.
Test Prompts & Configs
Copy-paste prompts and settings to verify your stack is running clean.
Full Walkthrough
Step-by-step screenshots and commands — start to finish, no gaps.
26B Parameters
Gemma 4 model size
17.99 GB
Download once, run forever
Port 1234
Local OpenAI-compatible API
macOS + Linux
16GB+ RAM recommended
Who It's For