The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
Title: AI Content Automation Pipeline (n8n)
Designed and built an automated content workflow using n8n and the Gemini API — a draft → humanize → quality-check pipeline that generates and refines content with minimal manual editing, cutting content production time significantly.
Tech: n8n, Gemini API, workflow automation
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
Awesome concept, super close to john wick!
What if an action movie had ZERO cuts?
I built a one-take chase sequence in Seedance 2.5; following a fugitive running through chaos, explosions, and close combat in real time.
The entire film is one continuous shot, designed to feel immersive, fast, and cinematic from start to...
The original was a bracelet on a grey concrete surface. The detail is all there, but nothing about it says "premium."
So I took the same piece and built a campaign around it. The first image shows the before and after. The next three are full ad concepts, each with its own setting, light, and mood, so you can see how one product can carry more than one story.
The bracelet stays the same in every frame. Only the world around it changes.
This is a personal concept project, not a client job. Brand names and trademarks belong to their owners.
Tools: Gemini,Nanobanana pro, GPT image 2.5
Worth naming the limit on that, though: the world around it is rentable. Any competitor can buy the same silk, the same rim light, the same stone plinth, so staging lifts perceived value right up to the point where two brands in a category stage the identical scene. The frame...
Most agent bugs are invisible: the agent loops, drops context or calls a tool it shouldn't, and all you see is a wrong answer. I wrote an agent harness from scratch in Python — loop control, context compaction, permissioned tools, sub-agents — and Harness Studio, a web app that replays every run: one lane per agent, each event explained, and the exact context each agent held when it stopped.
