How to turn voice notes into a newsletter
The newsletter you’re behind on doesn’t need more research. Most issues that never ship die the same way: you know what you want to say, you sit down to write it, and the writing session becomes a formatting-and-procrastinating session.
Talking fixes the blank page, but a raw transcript isn’t an issue either. Here’s the workflow in between.
The manual talk-first workflow
- Say the issue’s one idea out loud, in one sentence. If you can’t, you don’t have an issue yet, just a topic. Keep talking until the sentence shows up.
- Record 3–4 minutes answering, out loud: Why does this matter to my readers this week? What’s the story from my own work that shows it? What do I want them to do about it?
- Transcribe the recording.
- Mark the skeleton. In the transcript, mark the one-liner (that’s your subject line, roughly), the story, and the so-what. Delete everything unmarked — usually half.
- Rewrite the seams. Spoken transitions (“so anyway, the other thing…”) need one written sentence each. Everywhere else, keep the spoken phrasing, because that’s the voice your readers signed up for.
- One pass for cuts, then ship. Newsletter readers forgive imperfect prose; they unsubscribe from absent senders.
This is a genuinely good workflow. It takes 30-ish minutes once you’re practiced, and the issues sound like you because they started as you talking.
Where it leaks
Step 2 is an interview you’re running on yourself, and self-interviews go easy. You skip the question you don’t have a good answer to, and that is usually the question the issue needed. Then step 4 sends you back to editorial work you already did once out loud.
The same session, without the app-juggling
Run the same session in Xtraktr and the interview is real: you say the one idea, and it asks about the thing you just said — why this week, what happened, what’s the evidence, what’s the so-what. One question at a time, including the ones you’d skip. The skeleton step disappears because the Q&A structure is the skeleton: ask for an outline or a summary at the end and you’re at step 5, rewriting seams in your own words.
Ten minutes of talking, ten minutes of seam-work and cutting. That’s a shippable issue in about the time the manual version spends on transcription cleanup alone.
What makes it compound:
- Capture all week, don’t compose on deadline. A voice note after a client call or a session in the car is next week’s issue arriving early.
- Keep the half-formed stuff private by default. Your rough takes run through the AI on your own iPhone, and nothing you say reaches our servers. That means you can think out loud about a client situation or a competitor without pre-editing yourself.
