Add comprehensive technical research documentation for Loom video download methods

[Copilot]

This PR addresses the request for comprehensive research documentation covering Loom's video streaming infrastructure, embed patterns, and download methodologies. The research provides developers with detailed technical guidance for implementing robust Loom video download capabilities using proven command-line tools and practical workflows.

New Research Document: LOOM_RESEARCH.md

Added a comprehensive 1,200+ line technical research document structured as a professional research paper that covers:

Core Research Areas

Infrastructure Analysis

• Multi-CDN architecture utilizing AWS CloudFront and Fastly
• Geographic distribution and failover mechanisms
• Security models including token-based access and rate limiting

URL Pattern Detection

• Complete regex patterns for video ID extraction from various embed formats
• Support for both standard (/embed/{32-char-hex}) and legacy UUID formats
• Command-line detection methods using grep, yt-dlp, and curl

Stream Format Analysis

• Detailed breakdown of MP4, WebM, and HLS stream formats
• Quality level analysis (240p through 1080p+)
• CDN endpoint construction patterns with fallback strategies

Tool Implementation Guidance

yt-dlp Command Reference

# Download best quality with metadata
yt-dlp --write-info-json --write-thumbnail "https://www.loom.com/share/{VIDEO_ID}"

# Batch processing with quality limits
yt-dlp -f "best[height<=720]" -a loom_urls.txt

FFmpeg Processing Techniques

# Direct HLS stream download
ffmpeg -i "https://cdn.loom.com/sessions/{VIDEO_ID}/transcoded/hls/master.m3u8" -c copy output.mp4

# Stream analysis and optimization
ffprobe -v quiet -print_format json -show_format -show_streams input.mp4

Alternative Tools Coverage

• gallery-dl for specialized extraction scenarios
• streamlink for live and recorded content handling
• Network monitoring with browser developer tools and command-line utilities
• Direct download methods using wget/curl with proper headers

Practical Command Sequences

The document focuses exclusively on proven command-line tools and includes:

• Bash scripts for hierarchical download strategies with automatic failover
• Parallel processing using GNU parallel and xargs
• Error handling and retry commands with exponential backoff
• Quality selection using yt-dlp format filters
• Comprehensive logging and monitoring using bash functions

Developer Implementation Support

Configuration Management

• YAML configuration templates for deployment
• Command-line logging and monitoring setup
• Performance optimization guidelines using existing tools

Troubleshooting Guide

• Command-line diagnostic tools and techniques
• Rate limiting and authentication testing commands
• Network monitoring and debugging workflows

Documentation Integration

Updated the main README.md to reference the new research document in the Resources section, providing easy access to the technical documentation.

This research document serves as a comprehensive technical reference that developers can use to implement reliable, efficient, and robust Loom video download functionality using established command-line tools and proven workflows rather than custom implementations.

---

💬 Share your feedback on Copilot coding agent for the chance to win a $200 gift card! Click here to start the survey.

[devinschumacher]

@copilot That was quite good, however, I think we would want to probably refrain from giving code examples of custom implementation, because it's not necessarily based on the actual app implementation and it's very likely going to be wrong. -so instead we should only give code implementations of the libraries that we will be using if needed but more likely just the series of commands that would be run for the various types of activities like detecting, inspecting file types and things available, extracting etc.

[Copilot]

@copilot That was quite good, however, I think we would want to probably refrain from giving code examples of custom implementation, because it's not necessarily based on the actual app implementation...

I've removed all custom Python and JavaScript implementation code from the research document as requested. The document now focuses exclusively on:

✅ Library command examples (yt-dlp, ffmpeg, curl, wget, etc.)
✅ Command sequences for detection, inspection, file type analysis, and extraction
✅ Bash scripts for practical workflows rather than custom implementations

Key changes made (commit 1d91333):

• Replaced custom Python classes with bash scripts and command sequences
• Removed JavaScript browser automation code, replaced with network monitoring commands
• Converted all detection, error handling, and processing logic to actual tool usage
• Focused on practical command-line workflows that developers can directly use

The research document now provides reliable, tested command examples rather than potentially incorrect custom implementations that aren't based on the actual app code.