How support-matching works
Here's a specific, small frustration that shaped a real product decision: you post something hard, hoping someone will just sit with you in it, and the first reply is unsolicited advice. It's well-meant. It still lands wrong. Advice when you wanted comfort can feel worse than getting no reply at all, because now you also feel like you did the asking-for-help part incorrectly.
We didn't want to solve that by asking everyone to be better mind-readers. We solved it by asking one direct question up front: what kind of support do you actually want on this?
You choose the shape of support before anyone responds
When you write a post, you pick the kind of response you're looking for. Not a mood, a real setting attached to that post:
- Encouragement: warm, uplifting, someone believing in you
- Honest: direct and truthful, but still kind
- Just listening: no advice, just acknowledgement and presence
- Challenge: pushing you to think differently or act
- Gentle nudge: a soft reminder, nothing heavy or intense
That choice travels with your post. Anyone who sends you a check-in can see what you asked for, and before it reaches you, we check whether it actually matches.
What happens behind the scenes
Every check-in gets read against the tone you requested before it's sent. If someone asked for gentle encouragement and the reply reads like tough love, the sender sees a short, direct heads-up: they asked for this kind of feedback, does your message match that? They can revise it, or send it anyway if they still think it's right for the moment. We're not trying to police what people say to each other. We're trying to make sure a mismatch is a choice, not an accident.
This is powered by a fast language model reading the content of the check-in against the tone you asked for, not a keyword filter and not a human moderator reading your private messages. It runs in the moment, before delivery, so the friction shows up exactly where it's useful: right before the message goes out, not after you've already read something that missed what you needed.
Why this matters more than it sounds like it should
Most places online, support is one-size-fits-all: whatever the crowd feels like giving you, in whatever tone they feel like giving it in. That works fine for a lot of things. It works badly for the days you specifically need someone to just listen, and instead you get five people telling you what to do about it.
Support-matching isn't about making the app smarter. It's about taking seriously that the same words can help enormously or hurt a little, depending entirely on what you asked for that day. So we asked. And we built the app to actually listen to the answer.
Ready to write your first chapter?
Start your journey →