frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: WatchLLM – Debug AI agents step-by-step with cost attribution

1•Kaadz•1h ago
Hi HN! I built WatchLLM to solve two problems I kept hitting while building AI agents:

1. Debugging agents is painful - When your agent makes 20 tool calls and fails, good luck figuring out which decision was wrong. WatchLLM gives you a step-by-step timeline showing every decision, tool call, and model response with explanations for why the agent did what it did.

2. Agent costs spiral fast - Agents love getting stuck in loops or calling expensive tools repeatedly. WatchLLM tracks cost per step and flags anomalies like "loop detected - same action repeated 3x, wasted $0.012" or "high cost step - $0.08 exceeds threshold".

The core features:

Timeline view of every agent decision with cost breakdown Anomaly detection (loops, repeated tools, high-cost steps) Semantic caching that cuts 40-70% off your LLM bill as a bonus Works with OpenAI, Anthropic, Groq - just change your baseURL

It's built on ClickHouse for real-time telemetry and uses vector similarity for the caching layer. The agent debugger explains decisions using LLM-generated summaries of why each step happened. Right now it's free for up to 50K requests/month. I'm looking for early users who are building agents and want better observability into what's actually happening (and what it's costing). Try it: https://watchllm.dev Would love feedback on what other debugging features would be useful. What do you wish you had when your agents misbehave?

A Systems‑Level Architecture for Integrative Rejuvenation

https://forum.effectivealtruism.org/posts/aNmqSbya5FrG88eAf/a-systems-level-architecture-for-inte...
1•k_n_gk•2m ago•0 comments

Chinese Universities Surge in Global Rankings as U.S. Schools Slip

https://www.nytimes.com/2026/01/15/us/harvard-global-ranking-chinese-universities-trump-cuts.html
1•janpot•3m ago•0 comments

MapQuest: The Brief, Glorious Era of Printed Directions

https://multiverseemployeehandbook.com/blog/mapquest-the-brief-glorious-era-of-printed-directions/
1•TMEHpodcast•5m ago•0 comments

Ask HN: Does GitHub Copilot now leave unsolicited PR review comments?

1•blenderob•5m ago•0 comments

Why Kant and Sontag Cannot Speak Otherwise

https://jimiwen.substack.com/p/some-matters-on-taste
2•jimiwen•12m ago•0 comments

Vertical Solar Panels Survive Storms by 'Swaying' Like Trees

https://www.scientificamerican.com/article/vertical-solar-panels-wind-resistant-trackers-for-high...
1•sohkamyung•14m ago•0 comments

Cracking DXP and SXD

https://www.os2museum.com/wp/cracking-dxp-and-sxd/
1•ingve•14m ago•0 comments

Show HN: US Bank Statement Converter to Excel Ready for LLMs

https://usstatementconverter.com/
1•aleks5678•15m ago•0 comments

Show HN: Glot – Find internationalization issues in Next.js app

https://github.com/Sukitly/glotctl
1•sukit•15m ago•0 comments

Machado Presents Trump with Her Nobel Peace Prize Medal

https://www.nytimes.com/2026/01/15/world/americas/machado-trump-meeting-nobel-peace-prize.html
1•rootlocus•16m ago•0 comments

AI is just starting to change the legal profession

https://www.understandingai.org/p/ai-is-just-starting-to-change-the
2•s-macke•18m ago•1 comments

Bucketing optimization in SQL to deal with skewed data (BigQuery example)

https://smallbigdata.substack.com/p/bucketing-optimization-in-sql-to
1•tosh•21m ago•0 comments

What Was the Metaverse?

https://www.fastcompany.com/91467599/metaverse-zuckerberg-facebook-ai
1•agluszak•21m ago•0 comments

Building an agentic memory system for GitHub Copilot

https://github.blog/ai-and-ml/github-copilot/building-an-agentic-memory-system-for-github-copilot/
2•agluszak•22m ago•0 comments

World's Safest Airline Rankings for 2026

https://www.airlineratings.com/articles/worlds-safest-airlines-for-2026
2•austinallegro•24m ago•0 comments

Show HN: Native PyAnnote (speaker diarizer) in Rust

https://github.com/RustedBytes/pyannote-rs
1•yehors•25m ago•0 comments

Show HN: Automated tech news site with custom multi-LLM agent pipelines

https://wayr.today/how-it-works/
2•siddkgn•26m ago•2 comments

Just the Browser

https://justthebrowser.com/
3•cl3misch•26m ago•0 comments

McClatchy Media accuses Google in federal court of monopolizing online ad sales

https://www.miamiherald.com/news/business/article314325407.html
2•giuliomagnifico•28m ago•0 comments

Feldera's Visual Profiler

https://www.feldera.com/blog/introducing-feldera%27s-visual-profiler
2•lsuresh•33m ago•0 comments

UN treaty to protect 'extraordinary' marine life due to come into force

https://www.aljazeera.com/news/2026/1/16/un-treaty-to-protect-extraordinary-marine-life-due-to-co...
2•Qem•33m ago•0 comments

Show HN: Recursive Language Model for Querying Human Action by Ludwig von Mises

https://github.com/mateolafalce/human-action-rlm
1•lafalce•34m ago•0 comments

How Civilizations Fall: A Theory of Catabolic Collapse (2005) [pdf]

https://www.ecoshock.org/transcripts/greer_on_collapse.pdf
2•ColPanic•34m ago•1 comments

Simple data management for new prototypes

1•AndreyK1984•35m ago•0 comments

Show HN: HN Reader with Favorites, Read-Later and Open Source

https://hn-pb-next.mystack.host/en
1•lavren1974•35m ago•0 comments

Beyond Senior: Consider the staff path

https://hawksley.org/2026/01/14/beyond-senior.html
1•taubek•36m ago•0 comments

I created a game engine for Django?

https://en.andros.dev/blog/6e9e4485/i-created-a-game-engine-for-django/
1•andros•37m ago•0 comments

Tell HN: YouTube gave my username Switzerland to a half government organization

4•faebi•40m ago•1 comments

The Spectrum Between AI Agents and Workflows

https://www.webguideplus.com/2026/01/ai-agents-vs-workflows-why-llm-agents.html
1•Traumen•41m ago•1 comments

Show HN: Quint Visualizer – A GraphViz-like visualizer for Quint traces

https://quint-visualizer.noghartt.dev/
1•Noghartt•48m ago•0 comments