Reflct - The journaling habit you'll actually keep

@klara_minarikova 

Really good question and it's something I thought hard about while building this. Right now Reflct surfaces the pattern and stops there. So if your mood has been dropping for several weeks, it will name that clearly in your weekly summary or pattern detection. Something like "your mood has trended downward over the past three weeks, most noticeably mid-week."

The next step is intentionally left to you, the user. My reasoning: for data this personal, unsolicited advice can feel intrusive or even patronizing. "Your mood is low, here's what to do" assumes the AI knows your context, your life, your reasons. It doesn't. That said, I've had a few people suggest a softer version of this, not prescriptive advice but something like "others in similar patterns have found X helpful." That's a direction I'm genuinely considering.

The line I want to hold: Reflct should always feel like a mirror that occasionally points something out, never a coach telling you what to do.

Still figuring out exactly where that line sits.

Thanks for mentioning this 🙏

@linapok 

It's closer to acknowledgment than conversation. It reads the emotional weight of what you wrote and responds

to that, not just the content. Deliberately quiet, one short note, then lets you rest.

Still actively improving this part. But the goal was always to make you feel heard rather than start a dialogue.

@viktorgems Thanks for the idea!

I’ve thought about this a lot. Right now Reflct is built around reflection and acknowledgment rather than advice: the AI reflects back what it sees in your entries: themes, mood shifts, what connected your days, so you can notice patterns yourself. I’ve kept it that way so it feels like "being heard", not like a coach telling you what to do.

I’m still trying to figure out what people who journal actually want: whether most prefer just being heard, or also want some pointers and direction. If it turns out a lot of users expect something like that, I’d seriously consider adding it. At the same time, it feels like a sensitive area, if we lean on the AI for full-on advice, what if it sometimes steers someone the wrong way? So I’m balancing "should I add this?" with "how do I do it in a way that’s helpful and safe?".

Your feedback helps a lot, it’s something I’m actively thinking through. Thank you!

@minhajulll This is a genuinely great idea and honestly you're right that it fits the name perfectly. "On this day a year ago you wrote..." is something I've seen done poorly in other apps but the way you're describing it, tied to mood evolution and not just nostalgia, is actually meaningful. Seeing that you were anxious about something that no longer affects you at all is a powerful moment of perspective. Adding this to the roadmap. Thank you for the kind words.

@mordrag 

For AI features, Reflct uses Anthropic’s API. Only the data needed for each feature is sent (e.g. that week’s entries for your summary, or the last few weeks for patterns), never your full journal. Anthropic doesn’t use your content to train their models, and your entries aren’t read by humans for training or review (per their stated policies). Reflct doesn’t sell or share your data; you can export or delete it anytime from Settings. Full details are in the Privacy Policy.

I treat this as an ongoing priority and I’m open to ideas on how to improve safety and transparency when AI processes user input. If you have suggestions, I’d love to hear them.

Thank you!

@copywizard Thank you, this means a lot! The hero example point is valid and I'm revisiting that section this week. On prompt types and retention, Reflct is actually still learning this itself. Every question has a difficulty weight and every category builds a preference profile as users interact with them. Over time the app surfaces what resonates for each person specifically. Once there's enough data behind that I'll have a much more precise view on that.

Feel free to reach out at support@reflct.co, would be curious what you see on the messaging side too.

@lev_kerzhner I agree. Thank you very much!

@jscanzi 

These are exactly the right questions and honestly I don't have complete answers yet, it's too early for real retention data.

But here's my thinking on each: On exhausting 120+ questions, probably not the point. The variety isn't there to prevent repetition forever, it's there to prevent the "same question every day" fatigue that kills habit in week one. After that the habit either forms or it doesn't, independent of question novelty.

On reflection feeling like homework, well, this is the real problem you're naming and the one I think about most. The two minute constraint is an attempt to solve it but you're right that it's an attempt, not a proof. The acknowledgment at the end of each entry is the other piece, making it feel like something happened rather than just another task completed.

On whether AI patterns retain or just sell, I suspect you're right that patterns sell the signup. What actually retains is probably the acknowledgment loop and the streak. The patterns are a reason to come back after week three when the habit is fragile. Whether that's enough I'll know in about 30 days. The surprising insight point is where I think the real retention lives. Obvious feedback gets ignored. Something you didn't expect about yourself is worth coming back for. Ask me again in a month, I'll have real data by then.