🧠 DeepDaily

Learn anything deeply — one day at a time.

🌐 Live App: https://deepdaily.net/
🎥 Demo Video: https://www.youtube.com/watch?v=wp6DZ94k9Jo

DeepDaily helps you learn new topics without getting lost in scattered tabs, random tutorials, and inconsistent progress.

Enter any topic, choose your level, and DeepDaily generates a structured 7-day or 30-day learning path with daily lessons, curated resources, and progress tracking.

✅ What You Can Do

Generate a structured 7-day or 30-day learning path for any topic
Choose your level: beginner, intermediate, or advanced
Follow one lesson per day with clear progression
Get curated resources for each day
Track progress and resume where you left off
💬 Lesson Q&A
- Ask questions about the lesson of the current day
- Get focused follow-up explanations without leaving the learning flow

🎥 Demo

👉 Click to watch the full demo

📸 Product Preview

🏠 Generate a Learning Path

Turn any topic into a structured learning journey in seconds.

📅 Structured Daily Curriculum

Follow a clear, day-by-day roadmap with progress tracking and guided lessons.

🔍 Curated Resources & Embedded Content

Access high-quality resources, including articles, books, and embedded videos — all tailored to your learning path.

✨ Features

📚 AI-Generated Lessons
- Structured lessons generated in real-time using LLMs
- Clear explanations, examples, and exercises
🗺️ Structured Learning Paths
- Generate full 7-day or 30-day curricula
- Progressive difficulty and topic breakdown
- Resume exactly where you left off
🎯 Level-Adaptive Learning
- Choose your level: Beginner / Intermediate / Advanced
- Content dynamically adapts in depth and complexity
🔍 Smart Resource Discovery
- Automatically curated learning resources per day
- Domain scoring system (docs > tutorials > noise)
- High-quality links (React docs, Python docs, etc.)
🌐 Real Web Search Integration
- Uses live web search to enrich learning content
- Avoids hallucinated or outdated resources
💾 Save & Update Lessons
- Save generated lessons to your account
- Update existing lessons seamlessly
🔐 Authentication
- Firebase Authentication
- Secure user-based data storage
⚡ Streaming Experience
- Real-time lesson generation (SSE)
- Smooth UX while content is being created
🎯 Modern UI
- Built with TailwindCSS
- Dark mode support
- Fully responsive (mobile-first)

🤖 Agent Architecture

DeepDaily is powered by a system of custom-built AI agents developed from scratch, each responsible for a specific part of the learning pipeline.

These agents collaborate to transform a user-defined topic into a structured, level-adapted, multi-day learning experience.

🧠 Planner Agent

Designs the overall learning journey based on the topic and user level.

Responsibilities:

Breaks down a topic into a structured multi-day curriculum
Adapts roadmap complexity based on user level (beginner / intermediate / advanced)
Defines daily learning objectives
Ensures logical progression and concept coverage

📘 Lesson Generation Agent

Generates the core learning content for each day.

Responsibilities:

Creates clear, structured explanations for the given topic
Adapts depth and complexity to the user’s level
Adds examples, analogies, and practical context
Builds deep-dive sections for advanced understanding

🔍 Resource Discovery Agent

Enriches lessons with high-quality external knowledge using real web search.

Responsibilities:

Performs web search to find relevant learning materials
Scores sources using a custom domain-ranking system
Filters out noisy or low-value sources
Selects the most relevant resources based on topic and difficulty level

📅 Daily Curriculum Agent

Assembles each individual learning day into a cohesive experience.

Responsibilities:

Combines:
- Learning objectives
- Lesson content
- Resources
Ensures consistency across all days
Aligns content with topic scope and user level

✅ Evaluation & Refinement Agent

Ensures quality, clarity, and coherence across all generated content.

Responsibilities:

Reviews outputs for clarity and completeness
Detects gaps, inconsistencies, or redundancy
Refines content when needed

🔄 How It Works

User inputs a topic and selects a difficulty level
Planner Agent generates a structured multi-day roadmap
For each day:
- Lesson Agent generates content (streamed)
- Resource Agent fetches real-world resources
- Daily Agent assembles the final lesson
Evaluation Agent refines the output
Content is delivered in real-time to the user

Why DeepDaily Exists

A common problem with self-learning is not lack of information — it is lack of structure.

People often start with motivation, open too many tabs, consume disconnected resources, and lose momentum after a few days.

DeepDaily is built to reduce that friction by turning any topic into a focused daily learning path.

🚀 Why DeepDaily is Different

Unlike typical AI tools that generate one-off answers, DeepDaily:

🧠 Builds structured learning journeys, not just responses
🔄 Uses a multi-agent system, not a single prompt
🌐 Integrates real web knowledge, not just model memory
🎯 Adapts to user level dynamically
📅 Encourages daily consistency → long-term mastery

This is closer to an AI-powered “learning system” than a chatbot.

🏗️ Tech Stack

Frontend

Next.js 16+ (App Router)
React 19
TypeScript
TailwindCSS
Sonner

Backend

FastAPI
Python 3.11+
Async architecture
SQLAlchemy (async)
PostgreSQL / SQLite (dev)

AI Layer

Multi-LLM support (OpenAI, etc.)
Structured JSON outputs
Streaming responses (SSE)

Infrastructure

Vercel (Frontend)
Render (Backend)
Firebase Auth
Docker

🌍 Deployment

Frontend: Vercel
Backend: Render
Live URL: https://deepdaily.net/

📁 Project Structure

deepdaily/
│
├── apps/
│   ├── web/                # Next.js frontend
│   │   ├── app/
│   │   ├── components/
│   │   ├── lib/
│   │   └── types/
│   │
│   └── api/                # FastAPI backend
│       ├── app/
│       │   ├── services/
│       │   │   ├── agents/
│       │   │   ├── curriculum_service.py
│       │   │   └── ...
│       │   ├── routers/
│       │   ├── models/
│       │   └── schemas/
│       └── main.py

🚀 Getting Started

1. Clone the repo

git clone https://github.com/eliacharfe/deepdaily.git
cd deepdaily

2. Start the backend

cd apps/api
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
uvicorn app.main:app --reload

3. Start the frontend

cd apps/web
npm install
npm run dev

4. Configure environment variables

Create .env files for both backend and frontend.

Frontend (`apps/web/.env.local`)

NEXT_PUBLIC_API_URL=http://localhost:8000
NEXT_PUBLIC_API_BASE_URL=http://localhost:8000

NEXT_PUBLIC_FIREBASE_API_KEY=your_key
NEXT_PUBLIC_FIREBASE_AUTH_DOMAIN=your_project.firebaseapp.com
NEXT_PUBLIC_FIREBASE_PROJECT_ID=your_project_id
NEXT_PUBLIC_FIREBASE_STORAGE_BUCKET=your_project.appspot.com
NEXT_PUBLIC_FIREBASE_MESSAGING_SENDER_ID=your_sender_id
NEXT_PUBLIC_FIREBASE_APP_ID=your_app_id
NEXT_PUBLIC_FIREBASE_MEASUREMENT_ID=your_measurement_id

Backend (`apps/api/.env`)

APP_NAME=DeepDaily API
APP_ENV=development
APP_HOST=0.0.0.0
APP_PORT=8000
FRONTEND_URL=http://localhost:3000
FIREBASE_ADMIN_CREDENTIALS_PATH=./firebase.json
OPENAI_API_KEY=your_key
TAVILY_API_KEY=your_key
TAVILY_SEARCH_DEPTH=advanced
DATABASE_URL=your_database_url

💡 Future Improvements

Regenerate today's lesson/resources
Smarter review & retention system
Improved lesson Q&A memory
Better topic discovery ("Surprise me")
Mobile app (Flutter)

🧑‍💻 Author

Eliachar Feig

⭐ Contributing

Pull requests are welcome!

📄 License

MIT License

❤️ Philosophy

Consistent, structured daily learning compounds into mastery.

Try DeepDaily

Live app: https://deepdaily.net/
Demo: https://www.youtube.com/watch?v=wp6DZ94k9Jo

Name		Name	Last commit message	Last commit date
Latest commit History 145 Commits
apps		apps
docs/images		docs/images
.gitignore		.gitignore
README.md		README.md
package-lock.json		package-lock.json
package.json		package.json

Folders and files

Latest commit

History

Repository files navigation

🧠 DeepDaily

✅ What You Can Do

🎥 Demo

📸 Product Preview

🏠 Generate a Learning Path

📅 Structured Daily Curriculum

🔍 Curated Resources & Embedded Content

✨ Features

🤖 Agent Architecture

🧠 Planner Agent

📘 Lesson Generation Agent

🔍 Resource Discovery Agent

📅 Daily Curriculum Agent

✅ Evaluation & Refinement Agent

🔄 How It Works

Why DeepDaily Exists

🚀 Why DeepDaily is Different

🏗️ Tech Stack

Frontend

Backend

AI Layer

Infrastructure

🌍 Deployment

📁 Project Structure

🚀 Getting Started

1. Clone the repo

2. Start the backend

3. Start the frontend

4. Configure environment variables

Frontend (apps/web/.env.local)

Backend (apps/api/.env)

💡 Future Improvements

🧑‍💻 Author

⭐ Contributing

📄 License

❤️ Philosophy

Try DeepDaily

About

Topics

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Frontend (`apps/web/.env.local`)

Backend (`apps/api/.env`)

Packages