If you've ever used voice dictation for work, you know how quickly it gets difficult for real work communication and writing. You change your mind halfway through a sentence, and the transcriber types every word anyway. You go back to fix it, and sometimes you just give up and type the whole thing yourself.
Typing gives you a moment to think before you put words on the page. But voice dictation expects you to think out loud in perfect sentences, which almost nobody can actually do.
Scribe works differently. It feels more like talking to a smart writing assistant than dictating into a microphone. You give it the rough idea, the way you'd explain something to a friend, and it writes the actual message in your voice. You can ramble or change your mind and the message still comes out clean. The more you use it, the closer it gets to sounding like you.
A few things Scribe does:
1. Writes a full message from a rough spoken idea. Hit your Scribe hotkey and say "write a follow-up to John about the design review" and Willow drafts the email in your voice.
2. Replies in context. Inside an email or Slack thread, say "reply and tell him I'll send the deck by Friday" and Scribe uses the existing thread to write the response.
3. Rewrites any text you highlight. Press your Scribe hotkey on a selection and say what you want changed. Try "make this writing more clear" or "rewrite this in Mandarin."
4. Works directly inside the apps you already write in, with full awareness of the document or thread around you.
Close to 20% of ChatGPT usage is helping people with communication like drafting emails or rewriting messages. Scribe is built exactly for that job. It runs on your voice, lives inside the apps you're already using, and picks up your writing style.
Willow now has two voice modes. Dictation, our original product, lets you type with your voice anywhere on your computer. And Scribe writes the message for you when you'd rather speak the gist.
Between them, my hands have barely touched the keyboard today!
For the Product Hunt community, we're offering 50% off for the first three months. Link at the top.
Our team would like to hear your feedback. Try it for a few days, tell us what you'd want next, and we'll be in the thread answering questions all day.
<p>Great work on this launch! Running inside the apps people already use instead of asking them to change their workflow is exactly the right call</p>
<p>The interesting bit here is the gap between “how I talk while thinking” and “how I want the final message to read.” A lot of voice tools preserve the first too literally, which makes the output fast but still not quite sendable.</p><p></p><p>I’d be curious how Willow separates transient spoken scaffolding — false starts, caveats, messy ordering — from durable voice traits like sentence rhythm, level of warmth, and how direct someone tends to be. That feels especially important inside Slack/email, where the same person may want a very different register depending on recipient and risk.</p>
<p>The biggest issue for me has always been that I don’t speak in perfectly structured sentences, especially when writing emails, messages, or prompts. So the idea of turning a rough spoken thought into a clean message in your own voice feels much more useful than traditional dictation.<br></p><p>The context-aware replies are especially interesting. If it can understand the thread and draft something that actually fits the situation, that could save a lot of time. A few questions: how much does Willow learn from someone’s writing style over time, and can users control or reset that personalization? Also, how does pricing work after the Product Hunt discount?<br></p><p>Really curious to see where this goes. Good luck with the launch!</p>
