Field notes from HolaClaw
Essays, engineering deep-dives and short videos from the team — on building AI that belongs to you.
Latest posts
Qwen 3.8 27B on your Mac: 24 GB or 32 GB?
Qwen's new 27B agentic model comes in two sizes that matter on a Mac: the standard 4-bit build for 32 GB machines, and a 3-bit build that squeezes the same model into 24 GB. We tested both on MacBook Airs, measured their speed, and picked how much the model should think.
Aug 18 · 8 min readOpenClaw browser automation with Chrome
OpenClaw automations stall at captchas, 2FA, and bot walls. HolaClaw lets the agent drive the real Chrome on your Mac, in a separate profile, so sites trust it.
Aug 18 · 6 min readWhich local models can run OpenClaw on your Mac? We tested 6
We tested 6 open-weight models with OpenClaw on a Mac: what runs, what breaks, and the RAM you need. 24GB to experiment, 32GB for everyday use.
Aug 17 · 8 min readAutomatic backups for OpenClaw
HolaClaw now backs up your OpenClaw assistant, manually or on a schedule, so you can upgrade or change settings and roll back to a known-good state if something breaks.
Jul 30 · 4 min readAgent Browsers for OpenClaw
A quest to integrate browsers in OpenClaw, in an automated and resource-aware way
Jun 2 · 6 min readIntroducing HolaClaw!
A short tour of the out-of-the-box HolaClaw experience for people already building with it, or about to try it for the first time.
May 12 · 3 min readTutorials
Install OpenClaw on your Mac
The official installer end to end, plus the Ollama one-command and HolaClaw one-click routes, and how to keep it all running.
Use Muse Glimmer locally with OpenClaw
Serve Meta's 30B open model on your Mac with Ollama or Llama and point OpenClaw at it, or let HolaClaw handle it.
Use Qwen 3.8 27B locally with OpenClaw
Run Qwen 3.8 27B locally behind OpenClaw on your Mac: which quantization fits your RAM (24 GB vs 32 GB), how to tame its thinking mode, llama.cpp and Ollama setup, and the one-click route with HolaClaw.