anyone here using n8n for full youtube automation?

I’ve been using n8n in my workflow for a while now, mostly to generate images/videos and handle part of the editing depending on the channel.
for scripts and voice i still keep it a bit manual, feels more natural and seems to hold retention better.

wanted to ask if anyone here is running things fully automated with n8n, like from script all the way to upload without touching anything.

I run a few faceless channels and i’m trying to balance time vs quality. main question is if performance stays close to a more “handmade” channel or if it drops a lot.

if you’ve tested this properly, how did it go for you? any tips or things to avoid?

Honestly, full automation is a trap for most niches. If you go 100% n8n from script to upload, the quality usually hits a floor and YouTube eventually flags it as reused or low value. I tried it and the retention was garbage compared to when I spent even 20 mins tweaking the script manually. Maybe it works for brainrot shorts but for long form you definitely want that "handmade" feel for the hook at least. Keep the voice and script manual if you want to actually build something that lasts.

I am experimenting myself with n8n, I think it's a game changer but I don't fully master it yet.
If you have any tips feel free to share

Funnily enough, I created a bot for YouTube automation - and as per usual, my ADHD kicked in and after its pretty much complete, its gathering dust in Cyberspace. My idea was to allow users to download material for real life shorts/videos - i.e last pic before death and that kind of thing... Basically, the app allows users to tag their own material - AI then understands what the images/videos are and knows where to place them in "his story" generation - as per topic. Once I got my head around the rendering on Creatomate, it works pretty well.

Went full auto on n8n for 3 months, retention tanked to ~22% avg view duration, channel stopped getting pushed past 300 views per video. The bottleneck is always voice and pacing, not the stack — ElevenLabs helps but AI cadence still kills retention in story niches. Hybrid with a 10-minute human pass on the script and a quick voice re-record is still the money spot.

Fully automated n8n setups can work, but most successful faceless channels still keep some human input usually in scripting or final edits since pure automation often reduces quality and retention. The best results usually come from automating production tasks while manually shaping hooks and storytelling. Fully hands-off systems can scale, but they are less consistent and need frequent tweaking.