Type 'ok', Get a Full Reply: How Intent-to-Email Works
You can reply to a Gmail thread by typing a single word — "ok" — and word2mail turns it into a complete, correctly punctuated, contextually accurate email: greeting, body, sign-off. No commands to learn, no template to fill in. The system reads the thread you're replying to, treats your few characters as pure intent, not as a draft, and writes the actual email around it.
That's the whole mechanism in one paragraph. Here's how it actually works, and why it's a different claim than "voice typing for email."
Two steps, not one fused flow
word2mail splits the job into two separate steps. Step one: get text into the compose window — by speaking a full sentence (the extension transcribes it) or by typing anything at all, down to a two-word fragment. Step two: whatever text landed in that window gets turned into a complete, correctly formatted reply — greeting, body, sign-off.
Competing "speak to write" tools fuse those two steps into one: you dictate, and a draft appears in the same motion — dictation IS the draft engine. word2mail decouples them, and that's the whole point. Because step one doesn't require a full sentence, you can skip talking altogether. Type "ok", "no, Friday instead", or "sounds good but later" — no punctuation, no greeting, no capital letters — straight into the same window, and step two runs exactly the same way: word2mail reads the whole thread you have open, treats your fragment as your intent, and writes a full, correct reply around it.
The value here isn't "voice instead of typing." It's minimum effort, period — motor effort and cognitive effort both. You don't have to compose a sentence in your head before you produce it. You just have to know what you mean, in three words or fewer.
What "reading the thread" actually means
This only works because the extension already has the context: a content script running on mail.google.com reads the subject line, sender, and visible body of the thread open in your Gmail tab — the same text already on your screen. That thread text, plus your two-word intent, go to the AI model together. The model isn't guessing what you might want to say; it's completing a specific, visible conversation.
Nothing about this is a background scan of your inbox. word2mail doesn't request a Gmail API scope to bulk-read your mail, and it doesn't keep a copy of the thread once your reply comes back — the content exists only long enough to generate that one draft. You read and approve every reply before it sends.
A first-hand look
[PLACEHOLDER — First-hand measurement to insert before publish: run 20 real reply scenarios (one-word intents like "ok" / "no, Friday" / "thanks, will do") through word2mail, measure keystrokes and time-to-approved-draft against typing the same reply by hand, and screenshot one example end-to-end (thread → 3-word intent → generated draft). Do not publish placeholder numbers as if real — this section must be replaced with an actual measurement before this post goes live, per V2M content canon: no fabricated statistics.]
Why this framing matters more than "dictate your email"
Plenty of tools transcribe speech. Fewer tools turn three typed characters into a finished paragraph that actually answers what was asked. The gap between "I know what I want to say" and "I have produced a written sentence" is where most email procrastination lives — and it's exactly the gap intent-to-email removes. That's true whether the bottleneck is typing speed, a sore wrist, or just the twelve seconds of dread before opening a blank reply box.
Try it
word2mail adds a mic button and this intent field to Gmail. Free to start, no credit card required — install it, open a thread you've been avoiding, and reply with three words.
About the author: Piotr Kucharski builds and runs word2mail solo, from Poland, under "Piotr Kucharski Bioinformatyka." One developer, no support queue — questions land at admin@voicetomail.eu and a human (him) reads them.