Not every problem is JSON or JWT. A surprising amount of day-to-day engineering is still text: a regex that almost works, two config files that should be identical, Markdown that needs to become HTML, an identifier in the wrong case, a PR description that needs a word count. These five patterns cover most of it.
1. Debugging a regex before it ships
Regex bugs are silent. A pattern that "works" on the happy-path string fails on Unicode, newlines, or the one ID format production actually uses. Live match highlighting and capture groups beat console.log loops.
A Regex Tester with flags support is the fastest way to prove the pattern against real samples before it becomes a support ticket.
2. Diffing plain text (not only JSON)
Not everything is structured. Log excerpts, nginx configs, email templates, and prose docs need a line-oriented diff. When the files are JSON, prefer a structured JSON Diff Checker; for everything else, a Text Diff Checker highlights additions and deletions without pretending keys exist.
3. Markdown → HTML for previews and CMS paste
README drafts, changelog notes, and docs often start in Markdown and need to land as HTML in a CMS, email template, or static page. Doing that by hand introduces broken lists and unescaped entities.
A Markdown to HTML Converter gives you readable HTML instantly. When the HTML needs entities escaped for safe embedding, pair it with an HTML Entity Encoder.
4. Case conversion for identifiers and APIs
APIs want snake_case, TypeScript wants camelCase, CSS and URLs want kebab-case, and env vars want CONSTANT_CASE. Renaming by search-and-replace is how you miss the third occurrence in a protobuf field.
A Case Converter turns one paste into every common convention. For route segments and filenames specifically, a Slug Generator is the narrower tool that also lowercases and hyphenates safely.
5. Counting words and characters under constraints
PR templates, meta descriptions, tweet-length status pages, and LLM prompt budgets all care about length. Editors lie when soft wrapping and invisible characters are involved.
A Word & Character Counter that updates live is enough for most limits. When the constraint is tokens rather than words, switch to LLM token counting instead — word count and token count are not interchangeable.
The common thread
All five are local text transforms. There's rarely a reason to paste production logs, customer emails, or draft docs into a hosted "utility site" that logs requests. Client-side tools keep the feedback loop fast and the data on your machine.