iCeGaming

It's a plug-and-play HTTP orchestrator that sits between AI agents/clients and merchant backends, handling reliable cart → checkout → payment flows with built-in support for emerging standards:

- UCP discovery (GET /.well-known/ucp + capability manifest) - A2A envelopes (/api/v1/a2a/checkout, /a2a/cart endpoints) - AP2 payment intents (consent proofs, handler binding, strict mode) - MCP conformance references

v0.2.0 updates (just tagged): - Durable payment state survives restarts/pod crashes - Stricter AP2 validation (structured proofs + optional trusted issuers allowlist) - Prod hardening (required PUBLIC_BASE_URL, single-replica K8s defaults, persistence alignment) - Updated docs, release notes, CHANGELOG

Built in Rust (multi-crate workspace, Actix/Axum), Dockerized, full K8s manifests (Deployment/HPA/PDB/NetworkPolicy/etc.), observability, idempotency/outbox/dead-letters/reconciliation.

No Rust needed to use: env-var config + point agents at it. Ops-ready with mocks/examples.

Why now: Google/Shopify UCP push + AP2/ACPA2A patterns mean merchants/agent devs need a neutral layer for protocol chaos + reliability.

Very early (single contributor), but looking for: - Feedback on API/protocol handling/edge cases - Deployment experiences or perf thoughts - Early pilots/testers (happy to debug integration)

Thanks to anyone in Rust, agentic AI, or e-commerce protocols for input!

Hey HN,

This is a very early alpha prototype I built as a solo side project: Mortar Mayhem, an 8v8 multiplayer tactical mortar combat game running at 60 FPS entirely in the browser—no login, no install, just click and play.

Core loop: Join a team, aim mortars, lob bombs, cause massive chaos and destruction. It's super rough around the edges (bugs expected, balance nonexistent), but the goal right now is quick validation: does the core mayhem feel fun? Is the aiming/feedback satisfying? Too random/chaotic, or just right? What mechanics/features would make it more addictive?

Jump in here: https://mortar-mayhem.icegaming.org/

Feedback super welcome—positive, brutal, whatever. I'll be around to reply and iterate based on what people say.

Thanks for trying it out!

Quick update: I've now added Play vs Bots under Practice Mode!

Hey HN, I just moved the project to its permanent domain: what-i-paid.com (seamless thanks to @lovable.dev magic — no downtime).

The core idea: Tired of dynamic pricing games and no easy way to know what people actually pay for things (phones, rent, medical, cars, groceries, etc.). So I built a simple crowdsourced database:

Upload a photo of any receipt from anywhere in the world: that's all. AI reads it, extracts products/prices/categories automatically (no manual tagging needed). Adds to searchable trends, category pages, filters, etc.

It's early but growing — focused on transparency so consumers see real paid prices, not advertised ones. Tech-wise: Lovable handled the app transition smoothly. AI does the heavy lifting on receipt OCR + categorization. Would love feedback: Try uploading a recent receipt (groceries, travel, tech gadget ... whatever).

What surprises you in the trends? Any categories missing? Bugs in AI parsing? Happy to answer questions or iterate based on what you see.

Link: https://what-i-paid.com Thanks for looking!

Hey, happy it resonates. For the external calls, they're not monitored yet. We get visibility via:

shared HTTP client wrappers emitting OTel metrics

alerting on downstream service error spikes (often external root cause)

occasional shadow schema checks + CI drift detection for critical externals

Can be considered for the future though, glad you mentioned it.

A platform for content creators to collaborate with other creators and brands seamlessly. Everything from discovering opportunities, running campaigns, to getting paid is handled in one place. The MVP is live, fully functional, and handles payments. Feedback from creators, founders, and marketplace experts is welcome!

Totally agree. Doctors won’t leave their EHR systems or install new software.

That’s why Cloud Patient is entirely web based. No installs, no setup, just a browser. Even if doctors never touch it, patients can still submit their records, when/if they have them.

If a doctor wants to add a note and upload a test result, it takes seconds. The goal is to keep it so simple that using it feels effortless.

Hey, thanks for the comment! Yes, this is vibe coding for ease of validation with `lovable`. I built Cloud Patient out of frustration with how fragmented medical visit data is. I'm not in the healthcare industry, just someone who's experienced the pain of juggling multiple doctors, clinics, and systems where each one has its own portal, logins, and incomplete records.

The idea was to create a single, simple place where patients can see all their visit history, and doctors can add details without dealing with clunky hospital systems.

Right now, it's very early, more of a proof of concept, so attachments and some features are still being built. The long-term goal is to give patients control and clarity over their health records, rather than having it locked away in a dozen separate systems.

I live in Germany, and here it’s even harder to get access to your own health data. You go in for a visit, they run a bunch of tests, and unless it’s really urgent, like “you have one month to live” urgent, you basically hear nothing.

There’s no easy way to check if you should be watching your cholesterol, blood sugar, or anything else. The information just sits there, locked away, and you’re left guessing.

That’s a big part of why I wanted to build Cloud Patient, to make this kind of information accessible and transparent so patients aren’t left in the dark about their own health.

What it is Cloud Patient is a preview platform designed to bring all your doctor visit history into one place. Patients can view their visits, while doctors can securely add details for each appointment — notes, treatments, and attachments like lab results or scans.

The goal is to give patients a unified, accessible view of their health history without relying on fragmented clinic portals.

Why it matters

Patients often lack access to a full medical history, especially when they’ve seen multiple doctors or visited different hospitals.

Doctors and clinics use different systems, making it hard to share data seamlessly.

Having a single cloud-based place to store and review visits can:

Reduce duplicate tests and miscommunication.

Improve patient understanding of their health.

Enable smoother transfers of care between providers.

Current state (preview)

Working well:

Clean and intuitive UI.

Doctors can log visit details that instantly show up for patients.

Still in progress:

Attachments (lab results, scans, etc.) cannot be downloaded or viewed yet — this feature is coming soon.

Privacy, encryption, and compliance frameworks (GDPR/HIPAA) are still being finalized.

No external integrations yet (e.g., hospital systems, fitness apps).

Roadmap questions:

When will attachments be viewable and downloadable?

What encryption and security measures will protect sensitive health data?

Will there be multi-factor authentication for both patients and doctors?

How will integrations with existing hospital systems be handled (e.g., FHIR, HL7)?

What happens if a user wants to export all their data?

Bottom line Cloud Patient has the potential to simplify how people manage their medical history. If the team can deliver on privacy, compliance, and interoperability while making attachments fully functional, it could become a valuable tool for both patients and providers.

For now, it’s a solid preview with a promising vision.

Thanks for sharing! OfferZen is doing great work, but they mostly focus on South Africa and a few specific regions. My approach is more global, aiming to reverse the process entirely so companies apply directly to developers for full-time roles. The goal is to shift the responsibility to the hiring side and make things as simple as possible for developers

RPS Duel brings the classic game of Rock / Paper / Scissors to the modern web — but with a competitive twist.

You’re not just playing against a friend or a bot here — you’re battling real players across the globe in live matches, climbing a global ELO-powered leaderboard, and proving that quick reflexes and sharp instincts still matter in the simplest game of all time.

Here’s what’s inside:

Instant matchmaking: No waiting rooms — click Play, and you’re in.

Worldwide competition: Battle anyone, anywhere, anytime.

Ranked matches: Every win or loss affects your ELO rating.

Achievements (coming soon): Unlock badges for milestones and streaks.

No ads, no tracking: Just fast, pure fun.

The game is lightweight, runs right in your browser, and works great on both desktop and mobile. Whether you’re here to kill a few minutes or grind your way to the top spot, there’s always someone ready to duel.

Are you the ultimate Rock / Paper / Scissors champion? Jump in and prove it: https://rpsduel.com

Thanks! You’re right, Hired.com started with a similar idea but evolved into a different model over time.

What feels different today is that developers are more burnt out than ever from endless applications and automated rejections, while companies are actively competing for talent in a tighter market.

My goal is to keep the developer experience really lightweight, just a LinkedIn URL with no resumes or complex profiles, and have companies do the heavy lifting.

I appreciate the encouragement and will share progress as things evolve.

I built a small MVP that flips the traditional hiring process: instead of developers applying to jobs, companies apply directly to developers.

Developers create a free profile by pasting their LinkedIn URL — no resume uploads or endless form-filling. Companies pay a monthly subscription to browse the talent pool and send direct job applications to the developers they want to hire.

The idea is to make the experience less painful for developers and shift the effort (and cost) to the companies. There’s also a planned "Pro" tier for companies with extra features like analytics and hiring pipeline tools.

Here's a working preview: https://preview--pitch-forward.lovable.app/

Would love feedback:

Do you think this could actually help fix some of the broken hiring process?

What are the biggest pitfalls you see in reversing the model?

Just finished adding full animations for walk, idle, hurt, casting, and dying to all wizards! Now movement, taking damage, casting spells, and dying all have their own distinct effects. It feels way more alive and polished. Would love any feedback or suggestions for more improvements!

  Warlocks Update — Now with Custom Lobbies!

You can now create your own multiplayer lobbies in Warlocks and play with up to 8 friends! New in this update: Create & join custom lobbies Predefined spawn points in a safe-zone circle Live player list updates while waiting Match starts automatically when full No install — just open in browser: https://warlockjs.pages.dev/ Run two tabs to test locally. Feedback appreciated as always