Claude app · Designed & built · 2026
An AI message composer that sounds like you, not a chatbot
A local Claude app that turns raw notes into polished emails, social DMs, and comment replies. Eighteen tones, a human-writing layer that strips AI-isms, and nothing ever leaves your machine.
The problem
Most of a support or ops job is writing: replies, follow-ups, DMs, comment responses. AI can draft them, but the output usually reads like it came out of a chatbot, and pasting customer emails into a third-party tool raises a fair question about where that content ends up.
What it does
You pick a format, paste the message you’re replying to, drop in your rough thoughts as fragments or bullets, and choose a tone. It comes back with a ready-to-send message across three modes: email, social DMs, and comment replies.
What makes it different
- Private by design. It runs entirely on your own machine and calls the Anthropic API directly, so your API key and message content never pass through any third-party service.
- Eighteen tones across the three modes, from warm and empathetic to concise, formal, bold, or persuasive.
- A human-writing layer built into every prompt that strips out AI-isms before they reach the screen, so the result sounds like you wrote it with care.
- An interface that stays out of the way. Context on the left, your thoughts on the right, tone below, output underneath. Nothing to learn.
Why I built it
It’s the writing rules I follow every day, turned into a tool: lead with what’s been done, name the specific detail, close with a real question. The difference is that now they apply to every message, even at the end of a long shift.