About Local Tone
What Local Tone Does
Local Tone is an English drafting tool designed specifically for professionals who communicate daily in English for work and life. Unlike generic grammar checkers, it detects meaning-altering errors rooted in a user's native language habits and adapts the output to local business norms or casual speech.
It is not a language course. It is a dedicated writing pad that catches the native language writing habits that undermine professional credibility or social naturalness in a new country — things like over-hedging phrases from Mandarin politeness norms, verb-final sentence structures carried from Korean or Japanese, or the "please kindly" construction that reads as formal in many Asian business contexts but sounds odd to an Australian reader.
The tool supports six regional presets: Australian, Canadian, Irish, British, American, and Global English. Each preset applies different vocabulary, tone norms, and idiomatic conventions. A draft written for a Perth colleague should not read the same as one written for a Boston client — Local Tone handles that shift for you.
Why It Was Built
I built this tool to solve a problem I faced personally. I moved from Hong Kong to Perth, Western Australia — and despite having written in English professionally for years, I kept running into the same invisible friction. Emails that felt clear to me read as oddly formal or slightly off to Australian colleagues. Review feedback that mentioned "clarity" without specifics. A tenancy dispute email that I rewrote three times and still wasn't sure about.
Standard spell-checkers fix typos. Grammarly catches surface-level grammar. Neither addresses the deeper problem: the way your first language (L1) shapes the structure, formality, and tone of the English you write, in ways that are invisible to you but immediately noticeable to a native reader.
As a software engineer who had spent a decade building data platforms, I approached the problem the same way I would approach any pattern-recognition challenge. I wanted a system that could identify the specific phrasing patterns that flagged my writing as non-native to an Australian reader, explain why the rewrite was different so I could learn from it, and track those patterns over time so I could watch recurring habits shrink rather than correcting the same mistake indefinitely.
Local Tone is that tool. I use it myself for every non-trivial piece of writing I produce.
Who I Am
My name is Ray Choi. I am a software engineer with over a decade of professional experience in system design and large-scale data engineering. My career has spanned backend systems, data pipelines, and product development across technology companies in Hong Kong and Australia.
I relocated from Hong Kong to Perth, Western Australia, several years ago. That transition — professionally, culturally, and linguistically — is the direct origin of this product. I know what it feels like to be functionally fluent in English but still get edited, still get the "communication" comment in reviews, and still second-guess the tone of an email before sending it.
I have no formal linguistics training, but I have read extensively in applied linguistics, second-language acquisition, and language transfer research, and I have spent years writing professionally in a second language. The perspective I bring to Local Tone is that of an experienced practitioner, not a theorist.
Local Tone is a solo-built product. I am the only developer, the sole author of the articles in the Resources section, and the support contact. If you email me, I will reply.
How Local Tone Works Technically
Local Tone is a web application built with a React frontend and a Python FastAPI backend. It sends your draft to an LLM provider — Claude, Gemini, or GPT-4o — with a carefully constructed prompt that instructs the model to identify writing habits from your native language and produce a rewrite calibrated to your chosen region and tone mode.
The tool does not store your draft text after the session ends. It does store structured pattern metadata — the categories of errors identified, their frequency, and the dates — so the pattern dashboard can show trends over time. Free-tier users bring their own API key (BYOK) and pay the provider directly at roughly $0.01 per analysis. Paid tiers include a managed key with no setup required.
Editorial Standards
Every article in the Resources section is written by a named author — currently only me, Ray Choi — and reviewed against regional style guides before publication. Each article carries a published date and an updated date that is revised whenever the content changes.
I do not publish AI-generated text without a complete line-by-line rewrite. The articles reflect my actual experience navigating English in professional and everyday settings in Australia and internationally. Where I cite specific linguistic claims, I link to authoritative sources such as Macquarie Dictionary, Oxford style guides, or published applied linguistics research.
If you find an error, an outdated recommendation, or a claim you disagree with, please contact me. I take corrections seriously and update the articles accordingly.
Contact
If you have feedback, feature requests, or simply want to discuss the technical architecture, I welcome direct communication. You can reach me at [email protected]. I read every email and use them to continuously refine the product.
See the Contact page for more detail, including what to include in your message and expected response time.