← Back to Blog

Accuracy · 7 min read

How to Build a Personal Dictionary That Actually Sticks

How to Build a Personal Dictionary That Actually Sticks

How to Build a Personal Dictionary That Actually Sticks

I built my personal dictionary over three months. Here's the system that actually works , and the mistakes that don't.

I started with 10 terms. Three months later, I have 40. The dictionary catches mistakes before they happen, and I rarely correct the same word twice.

Here's the system I used to build it.

Start small (10 terms)

Don't try to add every term you use. Start with the 10 words you correct most often.

Mine were: Echo Flow, DataDab, SFSpeechRecognizer, CFBundleVersion, and six client names.

That's it. Ten terms. Add them, use the app for a week, and see what happens.

I resisted the urge to add everything at once. That patience paid off , I added terms that actually mattered, not words I thought might come up.

10
Terms to start with , add more only when you catch real mistakes

Feed it from real mistakes

Here's the key: don't guess what terms you'll need. Wait for mistakes, then add them.

Every time I correct a word, I add it to the dictionary. If I correct the same word twice in a week, it definitely goes in.

This keeps the dictionary grounded in reality. I'm not adding words I think I'll use , I'm adding words I actually use and actually get wrong.

Client names , protect the exact spelling
Product names , casing, spacing, punctuation
Technical terms , APIs, flags, environment variables
Acronyms , protect the exact form
Legal entities , "Inc." vs "Inc" matters

The monthly prune

Here's what most people skip: pruning.

Every month, I review my dictionary and remove terms I haven't used. If a client project ended and I'm not using those names anymore, they come out.

A bloated dictionary becomes a junk drawer. Terms you don't use anymore clutter the list and make it harder to find what matters.

I prune mine on the first of every month. It takes two minutes and keeps the dictionary tight.

Build and forget
Dictionary grows forever
  • Bloated with unused terms
  • Hard to find what matters
  • Junk drawer with search
Build and prune
Dictionary stays useful
  • Only active terms
  • Easy to find what matters
  • Tight and relevant

What NOT to add

Here's what I learned not to add:

Common words. If the dictionary gets "the" wrong, that's not a dictionary problem , that's a recognition problem.

Words you might use. Only add words you actually use and actually get wrong.

Every technical term in your field. Add the terms you use in your actual writing. Not every term you know.

Terms with uncertain spelling. If you're not sure how to spell it, look it up first. Adding a wrong spelling protects the wrong version.

The honest note

A personal dictionary is not a glossary. It's a mistake-prevention tool. Add terms you actually get wrong. Prune terms you no longer use. Keep it small and relevant.

The bottom line

Start with 10 terms. Feed it from real mistakes. Prune monthly.

That's the system. It's boring, and it works.

Try Echo Flow free for 14 days · How the Personal Dictionary works · Why accuracy fails on important words


Sources