BeforeWord
Your voice notes, turned into LinkedIn posts people read
8 followers
Your voice notes, turned into LinkedIn posts people read
8 followers
Speak a rough thought for a few minutes. BeforeWord listens, asks the one question that makes it better, and turns it into LinkedIn posts built on 1,787 patterns from posts that actually worked. It learns your voice over time. 7 days free.
Free Options
Launch Team
Framer AI AgentsDesign and publish professional sites with AI
Promoted
The voice learning part sounds really cool, but I'd love to see a way to save drafts or notes from the listening sessions themselves. Sometimes the best insights come mid-thought, and losing that raw material before it gets polished into a post feels like a missed chance.
@sonkupu12462 You're completely right - and here's the good news: we never actually throw the raw material away.
Every draft keeps the original transcript behind it; we just weren't showing it.
Fixed that because of your comment - it's live now: open any draft → "Your original words."
Comment to shipped in under an hour.
This is exactly why we launched here - thank you. 🙏
🎯 BeforeWord
Converting voice notes into posts sounds useful.
Have you found that users care more about sounding authentic or maximizing engagement?
I could see those leading to very different product directions.
@aryan787544 Genuinely good question, and we had to pick - you're right that they're different products.
We picked authentic, and honestly the users picked it for us.
When people edit their drafts, the instructions are almost always "less dramatic," "cut the inspirational ending," "that doesn't sound like me." In months of edit data, nobody has ever asked us to make a post more clickbaity. Not once.
Our take is that for executives the tradeoff mostly disappears anyway. Their audience isn't strangers, it's colleagues, investors and customers - people who know how they actually talk. One post that sounds fake costs more than one viral post earns so we want to stay away from the engagement maximizing product since the end game there will inevitably become "Write an entire post for me" so it just turns into AI slop.
That's the direction we're deliberately not taking. BeforeWord only works if you actually said something - it starts from your spoken words, keeps your best lines verbatim, and even shows you the raw transcript behind every draft. The moment a tool starts inventing thoughts for you, you're not building your reputation anymore, you're renting the machine's.
So the endgame we're betting on is the opposite one: the tool gets better at sounding like you, not better at replacing you. Curious which way you'd have gone though - you clearly think about this stuff.
the one-question approach is honestly clever, but it would be great if you could also flag which posts are predicted to underperform before I actually publish them, maybe with a quick confidence score based on the pattern match
@ahmet1wdx Thought hard about this one and I'll be straight: pre-publish scores would be fake precision - nobody can honestly predict a single post's reach, and I'd rather not sell astrology. What we do instead is track how your posts actually land after publishing and feed that back in (read the answer to Bünyamin's question above), so over time it learns things like "your story posts beat your opinion posts" - from your real results, not a guess (The app asks you how a post fared after a day or 2 and learns from that). Fair trade?
Dumped a rambling voice memo about a side project and the draft it pulled out actually sounded like me, which I wasn't expecting. The single follow-up question felt less pushy than most writing tools.
@ayenurakkulndd Its gets more like you as it learns. Glad it worked for you