Maintain - AI-Powered Content Recommendations

A personal content recommendation app that uses AI to suggest videos, articles, and podcasts based on your mood, interests, and preferences.

Features

✨ AI-Powered Recommendations - Get personalized content suggestions using Google Gemini AI
🎭 Mood-Based Filtering - Choose your mood (tired, curious, motivated, relaxed, bored, chill) for better matches
💾 Save & Organize - Save content to categorized lists (listen, learn, knowledge, tomorrow, other)
📜 History Tracking - View your past recommendations and revisit them anytime
🔥 Trending Content - See what's popular right now (optional YouTube integration)
🔍 Search & Filter - Find saved content easily with search and tag filters
🎨 Modern UI - Clean, responsive design with dark mode support
🎉 Toast Notifications - Get instant feedback on your actions
⚡ Error Boundaries - Graceful error handling for a stable experience

Tech Stack

• Next.js 15 with App Router
• React 19 for UI
• TypeScript for type safety
• Tailwind CSS 4 for styling
• Sonner for toast notifications
• Google Gemini API for AI recommendations
• YouTube API (optional) for trending content

Getting Started

Prerequisites

• Node.js 20+ or Bun
• pnpm (recommended), npm, or yarn
• Google Gemini API key (required)
• YouTube API key (optional, for trending content)

Installation

1. Clone the repository:

git clone <your-repo-url>
cd maintain-app

2. Install dependencies:

pnpm install

3. Create a .env.local file in the root directory:

GOOGLE_GEMINI_API_KEY=your_gemini_api_key_here
YOUTUBE_API_KEY=your_youtube_api_key_here  # Optional

4. Run the development server:

pnpm dev

5. Open http://localhost:3000 in your browser

How to Use

1. Set Up Your Profile

Go to Settings and add:

• Your name
• Hobbies and interests
• Programming languages you know
• Work context
• Favorite YouTubers

2. Get Recommendations

• Select your current mood
• Type what you're looking for (e.g., "40-minute coding tutorial")
• Get AI-powered suggestions tailored to you

3. Save Content

• Click the "🔖 Save" button on any suggestion
• Choose a category (listen, learn, knowledge, tomorrow, other)
• Access saved content in the Saves page

4. View History

• Go to History to see all your past recommendations
• Click to expand and view suggestions again
• Delete individual sessions or clear all history

5. Explore Trending

• On the main page, view trending content
• Based on YouTube's trending videos (if API key is set)
• Falls back to curated suggestions if unavailable

Project Structure

src/
├── app/
│   ├── (components)/     # React components
│   │   ├── Chat-client.tsx
│   │   ├── LoadingSkeleton.tsx
│   │   ├── SuggestionCard.tsx
│   │   └── SidebarLink.tsx
│   ├── api/              # API routes
│   │   ├── history/      # History management
│   │   ├── likes/        # Like functionality
│   │   ├── profile/      # Profile management
│   │   ├── recommend/    # AI recommendations
│   │   ├── saves/        # Save functionality
│   │   └── trending/     # Trending content
│   ├── history/          # History page
│   ├── saves/            # Saves page
│   ├── settings/         # Settings page
│   ├── error.tsx         # Error boundary
│   ├── global-error.tsx  # Global error handler
│   ├── layout.tsx        # Root layout
│   └── page.tsx          # Home page
├── lib/
│   ├── profile.ts        # Profile utilities
│   └── store.ts          # File-based storage
└── types.ts              # TypeScript types

API Keys

Google Gemini API

1. Go to Google AI Studio
2. Create a new API key
3. Add to .env.local as GOOGLE_GEMINI_API_KEY

YouTube API (Optional)

1. Go to Google Cloud Console
2. Create a project and enable YouTube Data API v3
3. Create credentials (API key)
4. Add to .env.local as YOUTUBE_API_KEY

Development

# Run development server
pnpm dev

# Build for production
pnpm build

# Start production server
pnpm start

# Run linting
pnpm lint

Features in Detail

AI Recommendations

• Uses Google Gemini 2.0 Flash model
• Considers your mood, profile, and message
• Fallback to saved content if AI is unavailable
• Parses JSON responses with tolerance for truncation

History Tracking

• Automatically saves all recommendation sessions
• Stores up to 50 recent sessions
• Shows relative time (e.g., "2h ago")
• Expandable to view full suggestions

Search & Filter

• Search by title, creator, or description
• Filter by multiple tags
• Sort by newest, oldest, or relevance
• Shows result count

Trending Content

• Fetches from YouTube API when available
• Falls back to curated suggestions
• Categorized by gaming, music, news, etc.
• Refreshes on page load

Contributing

This is a personal project, but suggestions and feedback are welcome!

License

MIT