GeeksforGeeks Course Downloader (down2.js)

why this project exists ?

Anyone would know how frustrating 🤯 it is when you are learning something 🧐 and your internet connection drops or is not stable, it takes away your flow of study 📚

An automated downloader that downloads ⬇️ course videos (Live and Recorded) directly from GeeksforGeeks.

Initial screen of the app

🚀 Features

• Session Caching: Saves your login status via auth.json to prevent repeated logins.
• User Inputs Caching: Remembers your previous preferences out of convenience (user_inputs.json).
• Headless Browser Navigation: Uses Playwright (Firefox) to simulate real human clicks.
• Network Interception: Captures master M3U8 playlists triggered by "Play" buttons.
• Multi-threaded Downloading: Offloads the heavy downloading and decrypting process to a background Worker Thread.
• Concurrent Segments: Downloads multiple .ts video chunks simultaneously using p-queue to speed up the process.

🛠️ Prerequisites

• Node.js installed
• Dependencies installed: npm install
• Playwright browsers installed: npx playwright install firefox

💻 Usage

Run the main script:

node down2.js

Interactive Prompts

Upon running, the script will ask you for:

1. Quality: Choose from highest, secondHighest, medium, lowest.
2. Simultaneous Downloads: Number of videos to download at the exact same time (e.g., 4).
3. Concurrency: Number of .ts segments to download concurrently per video (e.g., 20).
4. Course Name: The exact name of your course listed in your "My Courses" tab.
5. Primary Tab: "Live" or "Chapters".
6. Chapter Selection: Choose whether to download "All Chapters" or specify a range.

🧠 How it Works (Under the Hood)

The general architecture flows in 3 phases: Setup & Navigation, Scraping & Interception, and Queueing & Downloading.

1. Setup & Navigation (down2.js)

• Initialization: Loads cached inputs and asks for user preferences via inquirer.
• Authentication: Launches a Playwright Firefox browser. If auth.json is present, it uses that session. Otherwise, it prompts for a username and password, logs in, and saves the new session.
• Navigation: Clicks to "My Courses", searches for your course, navigates to the specified tab, and figures out how many chapters exist.

2. Scraping & Interception (utils/video_link_intercepter.js)

• Iterates over the selected range of chapters and recordings.
• For each recording, the script clicks the "Play" button.
• It attaches a network listener to the page to scan outgoing requests for .m3u8 or .mp4 URLs.
• It intercepts the master playlist URL, attaches the title/folder structure, and groups this metadata securely.

3. Queueing & Multi-threaded Downloads (Worker/worker.js)

• As URLs are successfully intercepted, they are added to an ObservableQueue.
• Worker Thread: A background worker is spun up Worker/worker.js.
• The worker uses p-queue to cap the number of active video downloads.
• Processing (utils/m3u8Downloader.js):
  • MP4: If the captured link is .mp4, it streams it directly.
  • M3U8 (HLS): It fetches the playlist file to find the available resolutions, selects the one matching your Quality preference, and parses the .ts segment list.
  • It fetches the AES decryption key (if applicable).
  • It downloads the .ts segments concurrently based on your Concurrency setting, decrypts them via crypto, and merges them into a final .ts (or .mp4) file on completion.

⚠️ Important Security Note

• The auth.json file generated contains your active GeeksforGeeks login session and live JWT tokens. It is ignored in .gitignore, but never share this file publicly.