Library

I want to build a daily Stoic practice web app called Amor del Fato. The idea is to blend classic Stoicism with modern self help, so someone can sit down each day, do a short guided session, and see how it actually changes them over time.

There should be a short first run flow that asks what they want to work on, like anger, fear, discipline, grief or purpose, then greets them by name and suggests a path. The home screen should show a time aware greeting, one clear next practice, quick actions for breathing, reflecting or a random pick, a daily reflection with a quote, their streak, and a peek at the last journal entry.

A session should walk through a check in on how they feel, an anchor quote, a self paced guided script, a silent timer with a breathing ring and a soft bell, a closing reflection, then a check out. Chart the before and after moods over time with a plain language weekly summary.

Include several guided courses, a few classic Stoic exercises, short philosopher bios, a journal, and a progress view with minutes and a day dot calendar.

Everything should stay in the browser with no account and no tracking, with a JSON export, and it should be installable and work offline. Keep it calm and warm, never gamified.

I want to build a static website that turns Wikipedia's list of unsolved problems in mathematics into something genuinely usable. It should pull every entry from that article automatically with a script, so the data is generated rather than hand typed, and store counts and fields consistently so the numbers shown on different pages always agree.

People should be able to browse all the problems, search across titles, statements, fields and solvers, filter by field, status and whether it's a Millennium Prize problem, and sort several ways, with all of that reflected in the URL so a view can be shared. Each problem gets a detail page with its statement, the intro of its own article, citations, and related problems with a reason given for each suggestion.

I also want people to track their own progress on problems and keep a private research journal with LaTeX notes, version history, and image attachments, all stored locally in the browser and encrypted with a passphrase that's never sent anywhere. Plus a lab with a handful of live interactive tools that actually compute the mathematics, like Collatz orbits, prime sieves and zeta function zeros, backed by tests. A glossary and quiz would round it out. No backend, no accounts, no tracking, works offline.

I want to build a public reference directory website for cemeteries around the world, built entirely from open data I can legally use. The facts should come from Wikidata, which is CC0, and images from Wikimedia Commons where each file's free license is individually verified, with visible credit next to every photo plus a full credits page listing them all.

It should be a static site with thousands of generated pages, one per cemetery, plus browsing by country and region, a search that works without a server, sitemaps, and RSS. Speed matters, so keep client side scripting basically nonexistent and process images into efficient sizes.

Each cemetery page should show its name, location, when it was established, and notable people buried there, but only public figures who have a recorded burial place, a death date, and an English Wikipedia article. Link out to Find a Grave and BillionGraves without copying anything from them. Only show opening hours if freshly fetched, with the date it was checked.

I also need data fetching scripts that cache raw API responses locally so I can rebuild the dataset anytime. Tone should stay respectful throughout.

I need a simple static landing page site for a family of open AI governance specifications called the Kinetic Gain Protocol Suite. It's the front door for around ten or eleven published specs spread across two regulated industries, so the page should introduce the suite, explain what the specs are for, and link out to each one plus the main repository.

Please keep it plain HTML and CSS with no build step at all, so I can preview it by just serving the folder locally. I'd like a dark hero at the top with an emerald and slate color scheme, clean readable sections below, and it should look right on a phone.

Round it out with the usual things a real site needs: favicons and app icons, a robots file, a sitemap, a changelog, an Apache 2.0 license, and a plain text file describing the site for AI crawlers. There should also be room for sub pages covering each vertical and some documentation.

Finally, set up an automatic deploy on push to the main branch that uploads the site over FTP to a subfolder on my host, using stored credentials.

I want to build a Kubernetes operator in Go that publishes governance documents to well-known URLs inside a cluster.

The idea is that instead of someone hand maintaining ConfigMaps, they write a single custom resource that says what kind of document it is, which well known path it belongs at, and the JSON payload itself. The operator validates that, then creates and keeps in sync a ConfigMap holding the payload keyed by the filename from the path, labelled by document type, and owned by the custom resource so it gets cleaned up when the resource is deleted. Status should report whether it's valid, which ConfigMap was published, and carry clear conditions so a bad type, bad path, or malformed JSON gets rejected loudly instead of silently shipping something broken.

Supported document types are an AEO doc, an agent card, a tool card, and a generic disclosure.

Please keep the validation and ConfigMap building logic pure and separate from the cluster code so it can be unit tested, and test the reconciler with a fake client covering publish, invalid spec, and missing resource cases. I also want a Helm chart that installs the CRD, a least privilege role, a non root service account and the manager deployment, plus a small distroless container build, a sample resource, a readme explaining the whole thing, and an AGPL license.

I want a Kubernetes operator in Go that turns monthly LLM spending limits into a cluster resource. I should be able to declare a budget object with a monthly USD limit, a warning threshold percentage that defaults to 80, and a pointer to a ConfigMap plus key where some other pipeline already writes the running total spend.

The controller should read that number, compare it to the limit, and report a state of UnderBudget, Warning, or OverBudget along with the percent used, plus raise Kubernetes events when a team crosses into warning or over budget. If the ConfigMap or key is missing, or the value is not a number, it should report Unknown with a clear failing condition instead of guessing. It should watch both the budget and the referenced ConfigMap so spend changes re-evaluate right away, with a periodic recheck as a safety net.

Please make `kubectl get budget` show name, limit, spend, used percent, and state in a nice table.

Keep the scoring logic pure and cluster free so it can be unit tested, and test the reconciler with a fake client. Include a Helm chart that installs the CRD, read only RBAC, a non root service account, and the manager, plus a distroless Dockerfile, a sample resource, and a README explaining the FinOps motivation. License it AGPL-3.0.

I need a marketing homepage for a payments and financial infrastructure company. The idea is that businesses of any size can use us to accept payments online and in person, run subscriptions and usage based billing, issue cards, move money across borders, and embed payments inside their own platforms.

Start with a bold hero explaining that we help businesses grow revenue, with a clear sign up button and a Google sign in option. Under that, show the main product areas in a visual way so someone can quickly see what we offer. Add a strip of big trust numbers like currencies supported, payments volume processed, uptime, and active subscriptions.

Then I want three audience sections: enterprises, startups, and software platforms, each with short copy plus customer stories that link to a case study and list the products used. Include a few quotes from customers, and a section about professional services, certified partners, and support plans.

Make it feel modern, clean, and polished with lots of white space and smooth scrolling. It should look great on phones too.

I want to build a community site where designers can show off their work and get discovered. The heart of it should be a big, visual homepage feed of design shots laid out in a grid of image cards, where each one shows the thumbnail, the designer's name and avatar, and simple like and view counts. Visitors should be able to browse without an account, click any shot to open a detail page with a larger image, a short description, tags, and comments.

I'd like people to be able to sign up, create a profile with their avatar, bio, and location, and upload their own shots. Following other designers and liking work should be part of it too, so a profile page shows someone's portfolio plus their followers.

Add search and filtering so people can browse by category or popularity, and a simple page for hiring designers. Keep the whole look clean, bright, and image first, with lots of white space so the work itself is the main thing you notice.