Ember
A compassionate, AI-powered companion for habit recovery
Start new thread

Ember - A compassionate, AI-powered companion for habit recovery

by•
Ember is a quiet, AI-powered behavior-change companion for people tired of rigid habit trackers. Built around psychological safety, Ember drops streak counters and guilt in favor of root-cause analysis. It dynamically maps your triggers, predicts high-risk windows, and delivers gentle, timely interventions right when you need them most.

Add a comment

Replies

Best
Maker
📌
Hey Product Hunt community! 👋 I built Ember to address a flaw in traditional habit trackers: they rely on streaks, badges, and guilt—which often backfire when real life happens. For anyone who has repeatedly tried and failed to break unwanted habits, consecutive streaks often create unnecessary pressure. When you miss a day, shame sets in, leading to abandonded apps. Ember is designed around "Quiet Persistence" and "Infinite Recovery": 🛡️ No Guilt, No Streaks: No streak counters, success percentages, or complex dashboards. 🔍 Pattern Before Prescription: Maps your specific triggers (fatigue, late nights, stress) and learns from relapses rather than punishing them. ⏱️ Predictive Nudges: Learns your high-risk windows and intervenes with micro-actions before you reach a breaking point. 🆘 Instant SOS Mode: A 1-tap grounding breathing exercise for high-urge moments. 🔒 Anonymous-First & Private: Local UUIDs, full data export, and 1-tap data deletion options. Ember is built with Vite, Supabase, and custom LLM summarization rules. You can try it out today on web or install it directly as a PWA at ember.yaabai.com. I’d love to hear your feedback, thoughts on the approach, or any questions about the tech! 🙌

The root-cause analysis part actually feels different from other apps I've tried, kind of like it gets why I keep slipping instead of just nagging me about it. Really gentle nudges too, not the usual preachy tone.

Maker

 Thanks so much Asya! You nailed the exact problem we set out to solve.

To keep things from feeling overwhelming, we use a few key principles:

  1. Micro-Actions First: Instead of overwhelming users with big tasks, suggestions are trimmed down to tiny, low-friction micro-actions (e.g., a 3-minute grounding exercise or a brief "Future Self Letter").

  2. Context-Aware Tag Matching: The AI matches suggestions directly to the user's specific trigger and mood tag (like Fatigue or Late Night) rather than dumping a massive library on them.

  3. Single Goal Focus: Ember intentionally limits focus to one primary goal at a time to keep cognitive load low when someone is already in a high-stress state.

Thanks again for the support! Would love to hear your thoughts if you get a chance to try it out.