frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Show HN: Frugon – Find which LLM calls a cheaper model could handle (local, MIT)

https://github.com/Rodiun/frugon
9•jarodrh•1d ago
I started leaning in on AI heavily this year, as I wanted to get more done autonomously, but then my token usage climbed dramatically to the point where my weekly quota would run out before the end of the week, sometimes a couple of days into the week.

I realised I had to do something about it else I'd have to double my spend. So I decided to start tracking my cost per task type. This revealed that a lot of my spend went to searches/scans or simple things like scouting tasks.

I then decided to turn this into a simple CLI tool that can be used to read your OpenAI-style logs locally, and analyze the cost and compare this spend to other models, then show you how much you could potentially save by switching those calls to a cheaper model.

When you run analyze you get an offline estimate priced against LiteLLM and gated by LMArena tiers. The general savings bands come from the research published by RouteLLM; but you can confirm this yourself using 2 commands --measure (shows the prompt-response output side by side) and --judge (a model chosen to do the comparisons). These send a sample of the prompts from the logs to the candidate models - either the default choice or set by you. This call goes directly to the model provider (never through me) as any normal LLM call would, and the response is shown and judged to either be better or worse or a tie.

It's deliberately small, because I tend to over complicate/think things sometimes: analyze + capture + a few commands, doing three jobs. Cost, quality visibility, routing recommendation.

Nothing is hosted. capture is an optional local proxy on your own machine, and there's no endpoint in the path of your data. You can confirm this by checking the source.

I included a demo so you can check out the output. It has a synthetic 56k call log (a month's worth) showing how costs can drop from $549.46 to $343.91 a month. A 37.4% saving.

Try it:

  uvx frugon analyze --demo
or

  uv tool install frugon
Then point it at your own logs.

All feedback is welcome, especially any on the routing/quality logic, or anything else, good or bad.

Comments

cyanydeez•44m ago
this would be more interesting as a local LLM anlysis; throw out all the costs, and figure out primary-subagent model architecture, and maximize token generation and prefill.

I don't see how anyone can operationalize this information.

XUEYANZ•17m ago
Haha. somehow i just love the naming. it just makes sense :D

John Deere owners will get the right to repair equipment under FTC settlement

https://apnews.com/article/john-deere-right-to-repair-agriculture-equipment-cb7514ffedb95c130a976...
263•djoldman•2h ago•56 comments

Separating signal from noise in coding evaluations

https://openai.com/index/separating-signal-from-noise-coding-evaluations/
154•sk4rekr0w•4h ago•59 comments

A software engineering interview question I like: computing the median

https://krisshamloo.com/blog/007
17•speckx•59m ago•7 comments

Chatto is now open source

https://www.hmans.dev/blog/chatto-is-open-source
728•speckx•10h ago•192 comments

Unicode's transliteration rules are Turing-complete

https://seriot.ch/computation/uts35/
31•beefburger•15h ago•4 comments

Mistral's Robostral Navigate: a state of the art robotics navigation model

https://mistral.ai/news/robostral-navigate/
412•ottomengis•11h ago•96 comments

Rewriting Bun in Rust

https://simonwillison.net/2026/Jul/8/rewriting-bun-in-rust/
5•doppp•15m ago•2 comments

Remote Attestation

https://www.liamcvw.com/p/remote-attestation
16•lcvw•1h ago•8 comments

Cloudflare Drop

https://www.cloudflare.com/drop/
242•coloneltcb•6h ago•125 comments

Show HN: Microsoft releases Flint, a visualization language for AI agents

https://microsoft.github.io/flint-chart/#/
198•chenglong-hn•7h ago•75 comments

Grok 4.5

https://x.ai/news/grok-4-5
469•BoumTAC•7h ago•581 comments

We made Grok 4.5, GPT-5.5, and Claude build the same apps

https://www.tryai.dev/blog/grok-4.5-vs-gpt-5.5-vs-claude-build-off
51•hershyb_•2h ago•19 comments

Turning a pile of documents into a searchable useable knowledge base

https://github.com/linuxrebel/DocuBrowser
73•linuxrebe1•5h ago•11 comments

Show HN: Yamanote.fun – A complete soundscape for Tokyo's Yamanote line

https://www.yamanote.fun/
52•madebymagnolia•1d ago•13 comments

MIRA: Multiplayer Interactive World Models Trained on Rocket League

https://mira-wm.com/
10•ethanlipson•1h ago•2 comments

FAANG Simulator

https://www.abeyk.com/escape-the-rat-race/
276•nerdbiscuits•5h ago•107 comments

GPT‑Live

https://openai.com/index/introducing-gpt-live/
600•logickkk1•8h ago•403 comments

A bug which affected only left handed users

https://shkspr.mobi/blog/2026/07/a-bug-which-only-affected-left-handed-users/
83•sixhobbits•12h ago•43 comments

Beyond Git: Real-Time Version Control for Godot – Lilith Duncan – GodotCon 2026 [video]

https://www.youtube.com/watch?v=CAJ_iIedx_I
10•surprisetalk•6d ago•1 comments

Rewriting Bun in Rust

https://bun.com/blog/bun-in-rust
274•afturner•3h ago•151 comments

Patching MechCommander's "left arm bug" for fun and profit

https://mhloppy.com/2026/05/mechcommander-weapons-left-arm-bug-fix/
5•Narann•3d ago•0 comments

TypeScript 7

https://devblogs.microsoft.com/typescript/announcing-typescript-7-0/
464•DanRosenwasser•9h ago•177 comments

New Sweden: the US's long-lost 'secret' colony

https://www.bbc.com/travel/article/20260629-new-sweden-the-uss-long-lost-secret-colony
46•bookofjoe•6h ago•3 comments

DKIM2 and DMARCbis Have Landed

https://stalw.art/blog/dkim2-dmarcbis/
75•StalwartLabs•2d ago•52 comments

Decoding the obfuscated bash script on a Uniqlo t-shirt

https://tris.sherliker.net/blog/obfuscated-self-evaluating-bash-script-by-cdn-akamai-being-suppli...
1297•speerer•16h ago•207 comments

OpenMandriva: Statement regarding attempted distribution sabotage

https://forum.openmandriva.org/t/statement-regarding-attempted-distribution-sabotage/8997
74•workethics•7h ago•13 comments

Show HN: Frugon – Find which LLM calls a cheaper model could handle (local, MIT)

https://github.com/Rodiun/frugon
9•jarodrh•1d ago•2 comments

Cloudflare Meerkat - Globally distributed consensus

https://blog.cloudflare.com/meerkat-introduction/
217•bobnamob•12h ago•45 comments

My road trip with the do-gooding cactus smugglers

https://economist.com/1843/2026/03/06/my-road-trip-with-the-do-gooding-cactus-smugglers
15•andsoitis•3d ago•1 comments

Open Source Barware: free, local-first bar inventory software (GPLv3)

https://opensourcebarware.com
21•RichBJamison•3h ago•11 comments