Content production pipeline · 2025
A publishing queue for more than one brand voice
ShippedIndependent engineer
Several brand voices published on a calendar held in one person's memory. Drafts moved by chat, and quality checks were inconsistent.
The constraint
The client keeps the voices. The pipeline keeps the queue. I do not produce the writing; the work is the queue and the checks.
Built with
Pythonn8nSEO
How it works
Intake into a queue. Draft routing by brand voice. SEO checks. Ready is a gate the pipeline enforces.
The calls that shaped it
Each decision with the pressure that forced it and the price it keeps costing.
Intake, route, check, ready
Drafts moved by chat and the quality checks were inconsistent, so ready meant whoever said it last.
"Ready" is a state produced by completed SEO checks that the pipeline enforces itself.
The cost: A draft cannot bypass the gate; publishing follows the checks.
n8n plus Python
Scheduling and hand-offs want a canvas an operator can see. The checks themselves want real code and tests.
Scheduling and hand-offs in n8n. Checks in Python. Neither is a chatbot with a calendar plugin.
The cost: Two systems to run, and the boundary between them is a maintenance line.
This is for teams that already have brand voices and a publishing calendar that keeps slipping. The writing stays with the authors; the queue and the checks are the work.
Where it stands
Shipped. Publishing is now driven by the queue, and the quality checks run on every piece.
- A publishing queue for a few brand voices, with the client keeping the writing.
- Ready is a state the pipeline produces from completed SEO checks.
- Scheduling and hand-offs run in n8n; the checks run in Python.
- Shipped.
What was handed over
- Queue states and who may move them
- Where the checks run
- How a brand voice is represented. The client still owns the voice.