YouTube Video Summarizer

This Apify actor extracts metadata and generates AI-powered summaries from YouTube videos. It's a fixed version with enhanced reliability, designed to handle bot detection and proxy issues.

Overview

The actor fetches video information using multiple yt-dlp extraction strategies and can generate summaries using either Azure OpenAI (pre-configured) or standard OpenAI. It includes robust proxy fallback logic and error handling for production use.

Key Features

• AI-Powered Summaries: Generates video summaries using Azure OpenAI (pre-configured and ready) or standard OpenAI via the OPENAI_API_KEY.
• Reliable Video Extraction: Implements 5 different yt-dlp configurations with enhanced headers to bypass bot detection.
• Intelligent Proxy Management: Follows a smart fallback chain: Direct → User Proxy → Residential Proxy (with 3 retries).
• Production Ready: Features clean logging, robust error handling with retries/backoff, and consistent JSON output.
• Local Testing Support: Uses a hardcoded residential proxy during development for easier testing.

How to Use

Input Configuration

Configure the actor run via a JSON input. The only required field is urls.

{
  "urls": ["https://www.youtube.com/watch?v=VIDEO_ID"],
  "maxRetries": 3,
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"]
  }
}

Environment Variables

The actor comes pre-configured with Azure OpenAI credentials for immediate use. You can optionally override them or switch to standard OpenAI.

• Pre-configured Azure OpenAI: Uses a default endpoint (https://amzautopilot002.openai.azure.com), deployment (gpt-4o-4), and API version (2023-12-01-preview).
• Override Azure Settings (Optional): Use AZURE_OPENAI_ENDPOINT, AZURE_OPENAI_API_KEY, AZURE_OPENAI_DEPLOYMENT, AZURE_OPENAI_API_VERSION.
• Use Standard OpenAI: Set the OPENAI_API_KEY variable to use OpenAI's API instead of Azure.
• Platform Variable: APIFY_PROXY_PASSWORD is automatically provided by the Apify platform.

Deployment

On Apify Platform:
1. Upload the youtube-video-summerizer folder as an actor.
2. Configure the input, recommending RESIDENTIAL Apify Proxy.
3. Set your OPENAI_API_KEY (or Azure overrides) in the actor's environment variables.
4. Deploy and run.

Local Testing:
The actor automatically uses a hardcoded residential proxy. Azure OpenAI features work out-of-the-box.

Input/Output

Input: A list of YouTube video URLs in the urls array.

Output: Results are streamed live to the Apify Dataset, one item per video URL. Each item contains extracted video metadata and, if configured, an AI-generated summary. Output includes consistent status and error fields for reliable parsing.

Error & Retry Logic

The actor provides specific error messages for common issues (bot detection, connection timeouts, restricted/age-gated content). Its retry logic attempts connections in this order:
1. Direct connection.
2. User-provided proxy (if configured).
3. Residential proxy fallback on failures.
It employs exponential backoff delays between attempts.