Hacker News: johalmedhttps://news.ycombinator.com/user?id=johalmedHacker News RSShttps://hnrss.org/hnrss v2.1.1Mon, 15 Jun 2026 10:40:19 +0000<![CDATA[LLM Budget Guard – open-source runtime cutoff for OpenAI/Anthropic]]>Article URL: https://www.llmeter.org/validate/budget-guard

Comments URL: https://news.ycombinator.com/item?id=47936677

Points: 1

# Comments: 0

]]>Tue, 28 Apr 2026 16:28:32 +0000https://www.llmeter.org/validate/budget-guardjohalmedhttps://news.ycombinator.com/item?id=47936677https://news.ycombinator.com/item?id=47936677<![CDATA[Sow HN: LLMeter – Track per-customer LLM costs across OpenAI, Anthropic,and more]]>Article URL: https://www.llmeter.org/

Comments URL: https://news.ycombinator.com/item?id=47653709

Points: 3

# Comments: 0

]]>Sun, 05 Apr 2026 20:48:01 +0000https://www.llmeter.org/johalmedhttps://news.ycombinator.com/item?id=47653709https://news.ycombinator.com/item?id=47653709<![CDATA[New comment by johalmed in "My API cost was at $13.19 when my persistent Claude named himself Thales"]]>I've been experimenting with routing requests between different models depending on the task complexity to keep costs down. It's surprising how much you can save just by defaulting to a smaller model for simple extraction tasks and reserving the heavy hitters only for complex reasoning. The tooling around this is definitely getting better, but tracking the actual spend per model in real-time is still a bit of a headache for side projects.

]]>Sun, 15 Mar 2026 00:49:02 +0000https://news.ycombinator.com/item?id=47383045johalmedhttps://news.ycombinator.com/item?id=47383045https://news.ycombinator.com/item?id=47383045