For press & media use. Last updated: March 2026.
Short answer: Subscription model. We charge for value delivered, not time spent.
Detail: Leda operates on a freemium model with paid subscriptions. Our success metric is whether users got what they needed — not how long they stayed. This is the same model as tools like Notion or 1Password: you pay because it's useful, not because it's addictive. Less usage per session actually reduces our infrastructure costs, so the incentives are aligned.
Short answer: Screen Time tells you what you did wrong. Leda gives you something better to do.
Detail: Screen Time and Digital Wellbeing are retrospective dashboards — they show you how much time you wasted after the fact. Leda is a proactive tool. Instead of shaming you for overuse, it provides focused mini-apps designed to deliver value in minutes, then nudges you to close the app when you're done. It replaces the doomscroll, rather than just measuring it.
Short answer: No push notifications, no streaks, no dark patterns, no engagement loops. The app actively nudges you to stop using it.
Detail: Anti-addictive design at Leda means we removed every mechanism that conventional apps use to pull users back: no push notifications by default, no streaks or gamification, no infinite scroll, no algorithmic feeds, no "you might also like" hooks. We also built active nudges — when Leda detects you've accomplished what you came for, it suggests you close the app. We have no investors pushing engagement metrics, because we're self-funded.
Short answer: Swiss startup, 11-person team, self-funded. No venture capital.
Detail: da.care SA is a Swiss-incorporated company (CHE-425.383.392) building ethical, anti-addictive AI products. The company is entirely self-funded — no VC, no investors pushing for growth-at-all-costs. The team of 11 is led by founder Philip, who has 20+ years of experience in consulting and technology ventures. Being self-funded means product decisions are driven by user wellbeing, not investor expectations.
Short answer: Multiple frontier models, intelligently routed. No single-model dependency.