Why this exists

Most developer utilities live in one of three places: a search result full of ads and interstitials, a desktop app that needs installing, or a one-off script in a repository. None of those is convenient when you need to decode a timestamp during an incident or check what a cron expression actually means.

This site is an attempt to make that lookup fast and predictable. Every tool has a page with an explanation, a working widget and the edge cases that the widget cannot show. The pages are static HTML, so they load without a login, a database or a background API.

How the tools work

Parsing, conversion and generation run in your browser. When you paste JSON into the formatter, the JSON is parsed by JavaScript on your machine. When you generate a UUID, the randomness comes from the browser's cryptographic random source. No tool sends its input to this site.

A few things still involve the network: loading the page itself, and any third-party advertising or analytics that may be enabled. Those requests are described in the privacy policy. The tool inputs are not part of them.

What is here

The collection currently contains 85 tool and reference pages across eight families. Each family has a hub page that explains the domain and links to the individual tools. The blog holds longer guides for the cases that need more than a widget: timezone transitions, the 2038 boundary, gateway errors, rebase safety.

  • Cron — generator, explainer, next-run calculator, syntax and Quartz references, common schedules.
  • Timestamp — epoch conversion, ISO 8601, timezones, relative dates, language snippets.
  • JSON — formatter, validator, diff, conversions to CSV, SQL, YAML, XML and TypeScript.
  • UUID — v4, v7, ULID, bulk generation, validation, timestamp decoding.
  • HTTP — a searchable status code reference with the commonly confused pairs explained.
  • Git — recipes for undo, rebase, cherry-pick, stash, bisect, submodules and branches.
  • Encoding — Base64, URL-safe Base64, data URIs, percent encoding, HTML entities.
  • Random — passwords, strings, numbers and hex colors.

Principles

No accounts. There is nothing to sign up for and no user data to manage. Closing the tab is the logout.

No hidden server work. If a tool needed a backend, it would have to say so. These do not. The tradeoff is that tools requiring large datasets, collaboration or persistence are out of scope.

Explain the edge case. A converter that silently guesses between seconds and milliseconds is worse than one that shows its assumption. Each page documents the assumption and the failure mode.

Static by design. The site is built with Next.js as a static export and deployed to a CDN. There is no runtime server to patch, and the pages are cheap to serve.

Corrections

If a tool gives a wrong result, the useful report includes the exact input, the expected output, the actual output and the browser. The contact page explains where to send it. Corrections to explanations and edge cases are welcome too.