Stop Rewriting the Same Dev Helpers: Build a Browser Toolbelt Instead

What changed for me ## A simple rule: refuse “one‑off” scripts for repeat work ## From random tabs to a deliberate toolbelt ## The core categories in my browser toolbelt ## 1. JSON and API helpers ## 2. Test data and fixtures ## 3. Time zones and date math ## 4. Encodings and “what is this string?” ## Keep it safe ## Where AI fits in (and where it doesn’t) ## When your toolbelt isn’t enough ## Build your own in 20 minutes ## Common pitfalls ## A tiny maintenance habit ## Over to you For a long time, my dev life looked like this: None of these tasks were hard: But they were constant. Death by a thousand tiny helpers. At some point I realised my problem wasn’t skill. It was that I treated every small task as a one‑off, instead of part of a repeatable workflow. So I stopped trying to be a hero and built myself a browser toolbelt instead. Once I treated these chores as a system, a few things shifted: None of this shows up on a project plan, but it shows up in how tired your brain feels at 3 p.m. This post walks through how I built that toolbelt, and how you can build your own with whatever tools you prefer. I did end up turning my own setup into a public toolbox at CodersTool, but the ideas here work even if you never visit the site. If a task is repetitive, mechanical, and easy to get wrong, I try not to solve it with a brand‑new script. Instead of opening a scratch file, I ask three questions: Often the answer is “no, I just need a reliable helper.” That’s where the browser toolbelt comes in. For years my “toolbelt” was just search history: Every session: same searches, slightly different sites, new pop‑ups and layouts. Two problems: So I did something boring but effective: That was the turning point. Once the folder existed, adding and pruning tools became natural. Your work will be different, but these categories gave me the best return. If you touch APIs, JSON is probably your main friction point. I rely on three things: You can cover this with: The key is to stop hunting and use the same few tools each time. I used to “test” with [email protected] and John Doe and be surprised when real data broke everything. Now I treat realistic test data as a first‑class tool: The workflow is simple: generate, export, import, re‑use. The more realistic your test data, the fewer weird production bugs you get. Time zones can eat entire afternoons. I mix: I never try to do timezone math in my head anymore. I push it through the same tiny set of tools every time. I keep one place for: You can do most of this with your standard library or openssl. I still like having a familiar browser tool for quick checks, especially when I’m away from my main machine. Browser tools are powerful, but they need guardrails: If it would be awkward to explain in a post‑mortem, it doesn’t belong in a public tool. Today we also have AI helpers like Copilot and ChatGPT. They are great for: But they are not a full replacement for a toolbelt: A small, deterministic toolbelt still wins for: I use AI around the toolbelt, not instead of it. Not everything belongs in a browser tool. Some signs you need a different approach: Those problems deserve proper code, tests, and review. For me, the toolbelt is for stateless, repetitive, mechanical tasks. If you want to try this, here is a quick starting plan: Start with three solid tools, not thirty. Add more only after you feel the same pain a few times. If you’re curious what my own browser toolbelt looks like in a more concrete, tool-by-tool way, the canonical, link-heavy version of this idea lives here: 👉 https://www.coderstool.com/blog/computer-programming-made-easier A few mistakes I ran into: Keep the bar simple: does this tool save me time, reliably, in real work? If not, it goes. Once a quarter, I spend ten minutes in the bookmark folder: That little cleanup keeps the toolbelt feeling light instead of bloated. That’s how I went from random tabs and throwaway helpers to a small browser toolbelt that quietly saves me hours a month. Share your essentials in the comments. I’d love to steal a few good ideas for my own setup. Templates let you quickly answer FAQs or store snippets for re-use. Are you sure you want to hide this comment? It will become hidden in your post, but will still be visible via the comment's permalink. Hide child comments as well For further actions, you may consider blocking this person and/or reporting abuse - Open editor. - Google “online JSON formatter” for the fifth time that week. - Write another tiny helper script I knew I had somewhere, but couldn’t find. - Format or validate JSON - Convert CSV ↔ JSON ↔ XML - Generate test data - Convert time zones - Decode some weird token or Base64 blob - I cut my “yak shaving” time by roughly 60–70%. - The little 5–10 minute detours started to disappear. - I went from many small context switches to a few predictable ones. - Have I solved this before? - Is there a stable tool that already does this? - Does this really need to live in the repo? - format json online - csv to json - utc to est converter - url decode online - Behaviour wasn’t consistent. - I never built muscle memory. - Picked one tool per repeated chore. - Bookmarked it in a Dev Toolbelt folder. - Used that same tool every time. - A formatter/validator so I can actually read payloads. - A way to diff two JSON blobs. - Converters between JSON and CSV or XML. - Your IDE formatter - CLI tools like jq - One trusted online formatter and converter - Faker‑style libraries for code - Simple generators for CSV / JSON lists - A shared sample dataset I can reuse - Language features and libs - MDN or official docs - One friendly converter for sanity checks - URL encoding/decoding - HTML entities - Quick hash checks - Never paste production secrets or live customer data into random sites. - Prefer tools that clearly say they process data in the browser. - For sensitive workflows, consider self‑hosted or in‑house tools. - Sketching one‑off scripts - Exploring unfamiliar APIs - Explaining confusing stack traces - AI can be confident but wrong. - It may use outdated or insecure patterns. - Results can change from run to run. - Quick format / convert / validate jobs - Reproducible workflows you can share with the team - Situations where you cannot paste data into an AI system - The task must run in CI or on every deploy. - The logic is specific to your business rules. - You need state, history, or heavy performance work. - List your top five recurring annoyances. - For each one, pick a single helper you like. - Create a Dev Toolbelt bookmark folder. - Add those links and pin the folder. - For one week, reach for the toolbelt before writing a new helper. - Add custom search keywords in your browser. - Pin a “toolbelt” tab you always keep open. - Share the list with your team and invite suggestions. - Trying to build a “perfect” toolbelt in one sitting - Re‑implementing tools that already exist and work well - Making everything depend on personal quirks no one else understands - Never pruning old or broken tools from the folder - Remove tools I no longer use. - Replace anything that feels slow or sketchy. - Add one or two helpers that proved useful in the last few months. - What is one tool you can’t live without in your daily dev work? - Do you keep a toolbelt like this, or is everything still hiding in search history and old repos?