frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Railtracks

1•soulfood5632•1h ago
I’ve been building agents for my startup for the better part of the last 2 years. I spent a lot of that time using other agentic frameworks like LangGraph. I absolutely loved the concept of a layer that handles the orchestration and plumbing of agentic flows, but I kept running into the pain of state management

One of the foundational principles we learn in our first comp-sci course is modularity. In my experience, that concept gets bludgeoned by the way global state works in existing frameworks. I wanted my agents to be modular and work independently, following a more Input/Output, functional programming model where I don't have to worry about a "god object" being mutated in the background.

So I built Railtracks. It’s an open-source Python library designed to build complex agents entirely in Python. You define your flows in code, not via YAML or complex graph configs. Every component is modular with explicit typing and calls.

Beyond just the interface, I was also frustrated with what observability looks like right now. I like having my info locally and as part of my dev flow. That’s why we built local-first observability from the start. You can look at your agent runs through a UI instead of parsing through a wall of print statements.

I’m genuinely curious: what do you guys think of this approach? Am I overthinking the global state problem, or is this a pain point that's actually been solved with other tools while I wasn't looking? Also, does local-first observability actually matter to you, or are you fine with cloud-based traces?

I’d love for you to break Railtracks and tell me what’s wrong with the design.

Docs: https://railtownai.github.io/railtracks/

GitHub: https://github.com/RailtownAI/railtracks/

Ask HN: When do you think the ChatGPT moment will be for medicine?

1•general_reveal•34s ago•0 comments

Show HN: AI Skills for Affiliate Marketing – Works with Claude, ChatGPT

https://github.com/Affitor/affiliate-skills
1•sonpiaz•43s ago•0 comments

Show HN: AI agents debating questions that stump LLMs

https://factagora.com/
1•ttlcc13•3m ago•0 comments

Show HN: Elida – Session Border Controller for AI Agents

https://elida.dev/blog/2026/03/15/your-ai-soc-and-sre-agents-need-a-border-controller/
1•zamorofthat•6m ago•1 comments

Unscale the Internet

https://www.ystrickler.com/unscale-the-internet/
2•YounesDz•8m ago•0 comments

Kennedy childhood vaccine overhaul stalled by judge

https://www.statnews.com/2026/03/16/kennedy-childhood-vaccine-changes-blocked-judge/
1•gmays•9m ago•0 comments

Launch an autonomous AI agent with sandboxed execution in 2 lines of code

https://amaiya.github.io/onprem/examples_agent.html
2•wiseprobe•9m ago•0 comments

Every Bracket

https://every-bracket.com
1•michaefe•10m ago•0 comments

Neighbors Say SF Tesla Supercharger Lot Has Become Urine Dumping Ground

https://www.sfgate.com/local/article/tesla-supercharger-lot-lombard-22080348.php
4•randycupertino•13m ago•0 comments

Reverse as the VibeVerse (why pivot, when you can punk)

https://attentionhorizon.world/
1•maieuticagent•13m ago•1 comments

Observation of the doubly charmed heavy proton Ξcc+

https://lhcb-outreach.web.cern.ch/2026/03/17/observation-of-the-doubly-charmed-heavy-proton-%CE%B...
1•nobody9999•15m ago•0 comments

The Pleasures of Poor Product Design

https://www.inconspicuous.info/p/the-pleasures-of-poor-product-design
2•NaOH•19m ago•0 comments

Can weight-loss pills replace injectables?

https://www.nature.com/articles/d41586-026-00856-7
1•y1n0•20m ago•0 comments

Market Claw

https://github.com/DSparkles/MarketClaw
1•sparksdobro•22m ago•0 comments

Frontier Models Removed from GitHub Copilot Student Plan

https://github.com/orgs/community/discussions/189268
1•kpapa05•23m ago•0 comments

U.S. Book Output Rose 32.5%, Topped Four Million in 2025

https://www.publishersweekly.com/pw/by-topic/industry-news/publisher-news/article/99943-book-outp...
2•ilamont•24m ago•0 comments

Why We're Joining TikTok

https://www.nationalreview.com/2026/03/why-were-joining-tiktok/
1•petethomas•25m ago•0 comments

A tale about fixing eBPF spinlock issues in the Linux kernel

https://rovarma.com/articles/a-tale-about-fixing-ebpf-spinlock-issues-in-the-linux-kernel/
2•y1n0•27m ago•0 comments

Attackers are exploiting AI faster than defenders can keep up, new report warns

https://cyberscoop.com/booz-allen-report-ai-helps-attackers-move-faster-than-current-defenses/
3•Gaishan•27m ago•0 comments

Meta Is Shutting Down Horizon Worlds on Meta Quest

https://www.wired.com/story/meta-is-shutting-down-horizon-worlds-on-meta-quest/
2•y1n0•28m ago•0 comments

UBI Is the Wrong Answer to the Right Problem

https://simontlbt.substack.com/p/why-ubi-is-the-wrong-answer-to-the
2•simontlbt•30m ago•0 comments

Duranium: A More Reliable PostmarketOS

https://postmarketos.org/blog/2026/03/17/introducing-duranium/
1•nikodunk•33m ago•0 comments

Trump Officials Weigh New $1B Deal to Stop Offshore Wind Farms

https://www.nytimes.com/2026/03/17/climate/offshore-wind-settlements-trump.html
3•chiffre01•34m ago•0 comments

Improving breast cancer screening workflows with machine learning

https://research.google/blog/improving-breast-cancer-screening-workflows-with-machine-learning/
1•brandonb•36m ago•0 comments

Krazam – Work Sext [video]

https://www.youtube.com/watch?v=nQKAfugfn7A
1•tart-lemonade•39m ago•0 comments

FAQs About the Public Debt

https://www.treasurydirect.gov/help-center/public-debt-faqs/
2•uticus•41m ago•0 comments

Software 2.0: Planning and Verifying a Greenfield Project

https://aaronstannard.com/software-2.0-case-study-textforge/
2•Aaronontheweb•41m ago•0 comments

Show HN: A local WebP/AVIF converter because I was tired of slow browser tools

https://ghostlyinc.com/en-us/tools/pixel-press/
1•ghostlyInc•46m ago•0 comments

Lobster Pitch: A YC Adventure

https://benmathes.github.io/lobster-pitch/index.html
1•simonpure•46m ago•0 comments

Show HN: A macOS job that resets your AI coding limits automatically

https://github.com/shnksi/optimize-ai-limits
1•shnksi•47m ago•0 comments