Hey everyone, my first post!
So, I’ve made a workflow in n8n that basically acts as a “text-to-comic” engine. It’s running on a stack of n8n + Google Gemini + Google Sheets + Drive.
Ling to gdrive with sample project (38 jpgs + 1 mp3 + 1 txt.
To be honest, this thing feels like a pocket-sized Hollywood studio. Watching it churn out content is wild, and I’m 100% sure there is a profitable business model hiding in here somewhere. But right now, I’m too deep in the dev side to see the best monetization angle.
How it works (Under the hood):
I broke it down into distinct agents to keep the AI from hallucinating and to maintain context. It’s not a “black box”—I’m running it with a human-in-the-loop approach for quality control.
-
The Director (Analysis Phase):
-
I feed it raw text (a poem, a rough script, whatever).
-
The first Gemini agent acts as a Director/Screenwriter. It breaks the text into scenes, decides the global setting (e.g., “Night Escapes,” “School”), defines the camera movement, lighting, and emotional goals.
-
It pushes structured JSON into Google Sheets (my backend) and creates a project folder on Drive.
-
-
The Artist (Prompt Engineering Magic):
-
This is where it gets cool. A second agent picks up the scene data.
-
Crucial Step: It pulls specific Character Descriptions from a separate database (Sheets). This ensures the main characters actually look consistent across panels, rather than changing faces every frame.
-
I hard-coded a specific style instruction (Jim Steranko vibe—dynamic, 60s spy aesthetic, bold colors). It generates 5-8 distinct, complex image prompts per scene.
-
-
The Render (Semi-Auto):
-
I keep a manual trigger here so I don’t burn credits on trash.
-
The workflow grabs the prompts, hits the Gemini Image Model, and generates the panels.
-
It auto-uploads the PNGs to the correct Drive folder and updates the Sheet with the file links.
-
Why this isn’t just another “wrapper”:
This isn’t just hitting a “Generate” button on Midjourney. It creates a coherent narrative structure.
-
It handles scene continuity.
-
It forces character consistency (mostly).
-
It manages the assets in folders automatically.
The Question:
I feel like I’m sitting on a content farm engine, but I’m stuck on the “Product” part. Where would you guys point this?
-
YouTube Automation? (Shorts with voiceovers over these panels?)
-
Tool for Indie Devs? (Generating assets for visual novels?)
-
Storyboard Service? (Selling pre-viz to ad agencies?)
The logic is solid, the code works. Just need to figure out how to package it. If anyone is interested, I can share more about how I handled the JSON structure to keep Gemini from breaking the format.
Give me some points of view pls )