AI-powered dating conversation analyzer. Upload or paste a chat screenshot to get brutally honest insights on attraction, ghosting risk, and what to say next.
Live: Deployed on GitHub Pages via CI/CD.
- Screenshot Analysis — Upload or paste PNG/JPG/WebP screenshots of dating app conversations. Supports multiple images at once.
- Sample Chat — One click runs the full analysis on a built-in demo conversation, no upload needed.
- Conversation Health Score — 0-100 rating of how balanced and engaging the conversation is.
- Attraction Score — Evidence-based assessment of how interested the match seems.
- Rizz Score — A separate read on your texting game, scored independently of how keen they are.
- Ghosting Risk — Predictive probability the match will stop responding.
- Brutal Verdict & Roast — A one-line headline verdict plus an honest jab at your own texting.
- Confidence Score — Flags when a chat was too short or too cropped to read reliably.
- Fake / Golddigger Detection — Risk assessment for fake profiles and transactional behavior.
- Reply Suggestions — 5 natural, tone-matched replies you can copy and send, regenerable in 8 different tones.
- Date Ideas — Casual, conversation-tailored date suggestions.
- Shareable Verdict — Copy a link that encodes only the verdict and scores. Chats, names and screenshots never leave your device.
- Angular 21 (standalone components, signals)
- Angular Material (Material 3 theming)
- TypeScript 5.9
- OpenAI Vision API (via Railway
ai-gatewaybackend proxy)
npm install
npm startOpens at http://localhost:4200/.
npm testnpm run buildBuild artifacts go to dist/DateSense/.
public/og-image.png is rendered from scripts/og-image.html — edit the HTML and
follow the regeneration command in its header comment to produce a new 1200x630 card.
Automatically deployed to GitHub Pages on push to main via .github/workflows/deploy.yml.