I Built a YouTube Deep Research Agent (Automates Research) скачать в хорошем качестве

I Built a YouTube Deep Research Agent (Automates Research)

In this video, I build a YouTube research agent using real multi-agent architecture. It analyzes top videos, finds content gaps, and generates new topic ideas automatically. Powered by LangGraph, TypeScript, and dynamic tool routing. Full breakdown + live test. Try it (please dont break my demo for the video thanks) https://www.nexforgeresearch.com/ [00:00:00] Intro: Deep research agent challenge [00:00:47] Multi-agent system overview [00:01:50] How the agent works: YouTube research process [00:03:00] Tech stack: Langraph and shared state [00:04:12] Orchestrator and delegation agents [00:05:02] Sub-agents and their roles [00:05:48] Context engineering and external storage [00:07:42] Embeddings and handling large transcripts [00:08:44] Agent setup: names, prompts, and recursion limits [00:09:38] Tool system and dynamic expansion [00:10:44] Side-by-side test: Claude vs. my agent [00:14:06] Research report generation process [00:15:51] My agent results and script creation [00:17:02] Lessons learned: clear problem, prompts, and tool descriptions [00:18:02] Business applications and future plans