Voice-First Writing: Capture, Structure, Polish, Send

The four modes

Each mode does one job. Capture generates material, structure orders it, polish fixes the language, and send ends it. Most stalls come from doing two at once, usually editing while speaking.

1. Capture: speak the whole thing, badly.
2. Structure: reorder, cut repeats, mark what is missing.
3. Polish: sentences, tone, grammar, names.
4. Send: ship it and close the document.

The same four modes cover a two-line reply and a long brief; only the time per mode changes, from seconds to an afternoon. What also changes is the device. Capture works anywhere you can talk. Structure and polish want specific tools, which is what the rest of this guide maps out.

Which tool for which mode

The modes stay the same across devices; the tools change, and so does which mode each device is good at.

If you move between devices, keep one home document per piece and paste every capture into it. Drafts scattered across three note apps go missing at the structure step, which is the step that needs all of them in one place.

Capture: talk past the mistakes

Success in capture is a complete rough draft, nothing else. Restarting a sentence because you disliked the wording rebuilds the blank page you were trying to escape.

• Keep talking through a stumble and let the wrong words stand. Auri's cleanup pass removes filler and false starts, and a plain spoken correction ("Tuesday, sorry, Wednesday") comes out as Wednesday.
• Mark gaps out loud: "stat goes here", "name to check". You search for those phrases during structure.
• Ignore tone. You fix tone in polish, with the whole draft in front of you.
• Stop when the idea exists on screen, not when it reads well.

Five to ten minutes is a full capture block. On Mac, the default shortcut behavior is Toggle: tap once to start, tap again to stop. If you think in bursts, switch to Hold under Settings > Shortcuts and Auri records only while the key is down. On iPhone, tap the Auri Keyboard mic and speak into the field itself, or open Smart Notes and record when the thought is longer than the field.

One capture through the modes

Here is one small piece moving through all four modes: a reply captured on a walk and finished at a desk.

You said: "reply to Dana about the launch, um, we can't do Friday because QA needs the weekend, say Monday, actually Monday morning, and ask if the deck is final"

Auri typed: "Reply to Dana about the launch. We can't do Friday because QA needs the weekend. Say Monday morning, and ask if the deck is final."

After structure: at the keyboard, the note-to-self line goes and the ask moves to the end. "We can't do Friday because QA needs the weekend. Monday morning works on our side. Is the deck final?"

After polish: with the draft selected, Edit Mode and "Make this more formal" produce the version that gets sent. "Hi Dana, Friday won't work on our side because QA needs the weekend. Could we aim for Monday morning? And is the deck final?"

Capture produced the material, cleanup made it readable, structure turned a note into a reply, and polish set the tone. Each step took seconds because each step had one job.

Structure: order before sentences

Structure saves the most time and gets skipped the most. Polishing paragraphs you are about to delete is the usual way to spend an hour and gain nothing.

1. Read the draft silently and break it into sections.
2. Move the real point to the top. When you talk, it usually arrives halfway down.
3. Delete the second and third versions of the same explanation.
4. Fill each spoken placeholder or delete the sentence around it.

Do this in text, with your hands, at the biggest screen you will be near that day. Reordering by voice is slower than typing; structure is the one mode where the keyboard clearly wins.

Polish, then send

Polish runs on a draft that already holds together. One pass, two at most; a third tends to trade real improvement for the feeling of progress.

• On Mac, select the draft, press the Edit Mode shortcut, and speak the instruction: "Fix grammar and spelling" for correctness, "Make this more formal" when the tone is off.
• On iPhone, open Auri Writing from the keyboard toolbar. Fix Grammar corrects the selection or the whole draft, Paraphrase rewrites it in one of more than 20 styles, and Translate covers 31 languages once the source is final.
• On either device, check names, numbers, dates, and anything you promised by hand. Cleanup and polish fix language, not facts.

For long-form work, polish deserves its own method: separate reads for structure, facts, and voice. The editing guide runs one full capture through all three.

Then send. Set the bar before you start: sendable, not perfect. Reread once, out loud if it matters, then send it and close the document so you stop reopening it. If one sentence keeps stalling you, say it out loud the plain way and type what you said. A draft that sits in a folder returns nothing on the capture, structure, and polish you put into it.

A day that survives interruptions

The workflow has to survive a day with three meetings and a delayed train. Short blocks in the right mode beat waiting for a free hour that never comes.

1. On the commute: capture two replies with the Auri Keyboard mic and one outline into Smart Notes. Do not structure anything on the train.
2. First desk block, twenty minutes: structure the outline on the Mac. Reorder, cut repeats, mark missing facts.
3. Before lunch: select the email that has been waiting, run Edit Mode with "Fix grammar and spelling", and send it.
4. Afternoon: a longer capture on Mac. Press the Auri Voice shortcut and talk the first draft straight into the document; pick a local engine first if the topic is sensitive.
5. End of day, ten minutes: one polish pass on whatever is closest to done, then send it.

When a meeting eats a block, leave the draft in whatever mode it reached. A structured outline waiting for polish is a far easier restart than a blank page, and the outline is still there tomorrow.

FAQ