AI-powered dating profile roast and optimization tool for Gen Z — brutally honest feedback, actionable glow-ups.
Most people on dating apps (Tinder, Hinge, Bumble) have no idea why they're not getting matches. Their friends are too polite to be useful, and generic advice articles are vague. The profile — bio, photos, prompts — is the product, and most users have never had it professionally critiqued.
Gen Z and Millennial dating app users (18–30) who are frustrated with low match rates and want honest, specific feedback on their profiles. Particularly users who have put in effort but aren't seeing results and suspect their presentation is the bottleneck.
The dating app market is massive and Gen Z expects AI-native experiences. A "roast" framing lowers the emotional barrier to receiving criticism — it's entertainment that happens to be useful. The key insight: people will pay for honest feedback they can't get from friends. The freemium model (free roast, paid glow-up) aligns with how this audience makes purchasing decisions — experience the value first, then convert.
- Roast framing — Critique as entertainment reduces defensiveness and increases shareability. A roast score people want to share is free distribution.
- Photo ranking as a standalone feature — Photo order on dating apps is critical and underappreciated. Ranking optimal photo order addresses a specific, high-impact pain point.
- Glow-up as the paid tier — Free roast = hook. Paid glow-up (rewritten bio, optimized prompts) = the actual value users want after they understand the problem. One-time payment model reduces friction.
- Google Gemini 1.5 Flash — Chosen for speed and cost. Profile analysis needs to feel instant; Gemini Flash delivers sub-3s responses at low cost at this volume.
- Factor-based scoring — Replaced rigid rigid scoring bands with a continuous factor-based calculation to avoid cliff-edge score jumps that feel unfair.
- Freemium gates removed at launch — Temporarily disabled payment gates to maximize early user acquisition and collect real usage data before re-enabling monetization.
User uploads photos + bio/prompts
↓
Next.js 14 App Router (Supabase Auth)
↓
Google Gemini 1.5 Flash
├── Profile analysis (tone, specificity, red flags)
├── Photo ranking (composition, expression, context scoring)
└── Glow-up generation (bio rewrite, prompt rewrites)
↓
Factor-based scoring engine (lib/)
↓
Results page (roast + score + ranked photos)
↓
[Paid] Glow-up delivery via Dodo Payments unlock
↓
Supabase (session, submission storage)
Stack: Next.js 14 · TypeScript · Supabase · Google Gemini 1.5 Flash · Dodo Payments · Vercel · Tailwind CSS
| Metric | Target (First 30 Days) |
|---|---|
| Sign-ups | 100+ |
| Paid conversions | ≥5% |
| Profile shares | 500+ |
| Avg. session time | >3 min |
| Roast completion rate | >80% |
- Re-enable payments — reactivate Dodo Payments integration after validating product-market fit with free users
- Platform-specific tips — tailor advice for Tinder vs. Hinge vs. Bumble (different mechanics, different optimal profiles)
- Before/after comparison — let users A/B test original vs. optimized profile and report back on match rate delta
- Group roast mode — submit friend's profile for a roast (viral loop)
- Video profile support — Hinge and others are adding video; expand analysis beyond static photos
- Streak tracking — weekly profile check-ins to track improvement over time