Hacker News: CDieumegardhttps://news.ycombinator.com/user?id=CDieumegardHacker News RSShttps://hnrss.org/hnrss v2.1.1Fri, 17 Jul 2026 01:09:41 +0000<![CDATA[New comment by CDieumegard in "Show HN: LocalClaw – Find the right local LLM for your exact hardware"]]>I'm currently building a full MAC application to install HomeBrew, Node, LM Studio, an adapted LLM, and OpenClaw locally.

WIP V1 is done, but I want a one-click solution

]]>Sat, 21 Feb 2026 12:06:44 +0000https://news.ycombinator.com/item?id=47100014CDieumegardhttps://news.ycombinator.com/item?id=47100014https://news.ycombinator.com/item?id=47100014<![CDATA[New comment by CDieumegard in "Show HN: LocalClaw – Find the right local LLM for your exact hardware"]]>Major UI update on localclaw.io + Dark/Light version added

]]>Sun, 15 Feb 2026 20:27:51 +0000https://news.ycombinator.com/item?id=47027220CDieumegardhttps://news.ycombinator.com/item?id=47027220https://news.ycombinator.com/item?id=47027220<![CDATA[New comment by CDieumegard in "Show HN: LocalClaw – Find the right local LLM for your exact hardware"]]>I built this because I kept seeing the same question everywhere / "what model should I run on my 16GB Mac?" or "will this fit on my 4060?"

LocalClaw matches your exact hardware (OS, RAM, GPU VRAM) to the right model + quantization out of 125+ options. Takes about 30 seconds.

It recommends for LM Studio specifically, shows file sizes, speed/quality scores, and explains WHY a model fits your setup (RAM usage %, VRAM fit, context window overhead).

Everything runs in the browser. No data collected, no account needed.

Built this in Switzerland, feedback welcome /|/ especially if a recommendation feels off for your hardware.

]]>Fri, 13 Feb 2026 07:23:58 +0000https://news.ycombinator.com/item?id=46999896CDieumegardhttps://news.ycombinator.com/item?id=46999896https://news.ycombinator.com/item?id=46999896<![CDATA[Show HN: LocalClaw – Find the right local LLM for your exact hardware]]>Article URL: https://localclaw.io/

Comments URL: https://news.ycombinator.com/item?id=46999891

Points: 1

# Comments: 3

]]>Fri, 13 Feb 2026 07:23:19 +0000https://localclaw.io/CDieumegardhttps://news.ycombinator.com/item?id=46999891https://news.ycombinator.com/item?id=46999891<![CDATA[New comment by CDieumegard in "Warcraft III Peon Voice Notifications but for Codex"]]>Hahaha soo much fun ! Congrats! :)

]]>Thu, 12 Feb 2026 08:53:43 +0000https://news.ycombinator.com/item?id=46986389CDieumegardhttps://news.ycombinator.com/item?id=46986389https://news.ycombinator.com/item?id=46986389<![CDATA[New comment by CDieumegard in "GLM-5: Targeting complex systems engineering and long-horizon agentic tasks"]]>Interesting timing — GLM-4.7 was already impressive for local use on 24GB+ setups. Curious to see when the distilled/quantized versions of GLM-5 drop. The gap between what you can run via API vs locally keeps shrinking. I've been tracking which models actually run well at each RAM tier and the Chinese models (Qwen, DeepSeek, GLM) are dominating the local inference space right now

]]>Wed, 11 Feb 2026 22:14:06 +0000https://news.ycombinator.com/item?id=46981961CDieumegardhttps://news.ycombinator.com/item?id=46981961https://news.ycombinator.com/item?id=46981961