Descriptions:
I built Nova, an AI agent whose only job is to find YouTube videos worth filming.
Instead of just generating random titles, Nova checks my posted videos, Notion pipeline, approved and rejected ideas, competitor videos, vidIQ signals, source transcripts, and channel history before it recommends anything. The goal is simple: stop wasting filming days on weak ideas.
In this video, I show how Nova works, why AI YouTube automation usually misses the real leverage, and how you can install the same agent for your own channel.
Get Nova here:
https://github.com/sharbelxyz/nova-youtube-agent
Use this prompt with your coding agent:
Install and set up this repo for me: https://github.com/sharbelxyz/nova-youtube-agent
Read the README and codebase first, then walk me through the setup step by step. Before changing anything, tell me what this repo does, what dependencies it needs, what accounts or APIs I need, what files I need to configure, how to run it locally, and how to test that it works after setup. Help me connect it to my own YouTube workflow.
Timestamps:
0:00 I built an AI agent for viral YouTube videos
0:34 Why AI YouTube automation misses the real leverage
1:35 Can AI choose my next filming slate?
2:06 Nova’s job: strongest ideas with receipts
2:36 The data Nova checks before recommending
3:06 Why evidence labels matter
3:37 The full Nova research sequence
4:37 Finding the viral pond
5:07 Reading source transcripts and adapting patterns
5:37 Scoring ideas for demand, proof, and freshness
6:41 How to install Nova for your own channel
7:15 The exact setup prompt
8:18 Connecting Nova to your workflow
9:19 Why real channel context matters
9:51 The weekly prompt to use
10:52 Live Nova demo inside Hermes
11:24 Nova reads my files and vidIQ data
11:56 The two video ideas Nova recommended
13:27 The real leverage in YouTube strategy
14:30 Comment Nova
The best AI YouTube system is not the one that makes videos for you. It is the one that helps you choose the right videos before you waste an entire filming day.
