Your personal AI-powered reading assistant — save with intent, get summaries, and organize your reading life.

• Dashboard: https://laterly-ai.vercel.app/ (highlight)
• API Base: https://laterly-ai.onrender.com/api/

Laterly.ai combines a Chrome extension and a web dashboard to help you capture articles with context, summarize them, and resurface what matters.

• 🔖 Save any page from the extension or manually from the dashboard
• 💬 Add “why am I saving this?” intent to keep context
• 🧠 AI-generated summaries + topic tagging
• 🔎 Search by intent, title, topic, or text
• 📊 Dashboard to review, bookmark, and mark as read

• Chrome Extension (MV3): one-click save with optional intent; auto-grabs page content.
• Manual Save (Dashboard): enter URL, title, intent, and content without the extension.
• AI Summaries & Topics: content is sent to AI for summary, topic detection, and intent normalization.
• Smart Search: filter by intent or topics; view trending intents.
• Read/Bookmark States: manage your queue from the dashboard.

• Node.js (v20 recommended)
• npm
• Chrome (for loading the extension)

# extension
cd extension && npm install

# dashboard
cd ../dashboard && npm install

# backend
cd ../backend && npm install

• extension/.env uses VITE_DASHBOARD_URL and VITE_BACKEND_URL.
• dashboard/.env expects VITE_API_URL pointing to the backend (e.g., https://laterly-ai.onrender.com).
• backend/.env should include DB URL, JWT secret, and OpenAI creds (see sample in repo).

cd extension
npm run dev

Load the built dist/ folder in chrome://extensions as an unpacked extension (Manifest v3).

cd dashboard
npm run dev

Open http://localhost:5173

cd backend
npm run start:dev

Backend serves under /api (e.g., http://localhost:3000/api).

• Save with intent via extension
• Manual save from dashboard
• AI summaries + topic detection
• Dashboard search + trending intents
• Weekly digest emails
• Topic-based exploration UX
• Shared/collaborative reading lists

PRs welcome! If you're passionate about productivity, AI, or reading — open an issue.

Built by @SharadSaha

MIT © 2025