CONTENT OPERATIONS
SEO and Blog Automation
A content pipeline running from topic discovery and research through drafting, imagery and WordPress publishing, with an approval step before anything goes live.
- PROJECT
- SEO and Blog Automation
- INDUSTRY
- Content Operations
- STACK
- n8n · Google Gemini · Tavily · WordPress REST API · Google Sheets · AI Image Generation
OVERVIEW
The project
Publishing consistently is an operations problem more than a writing one. Allzone built a pipeline that generates topics aligned to a service list, researches each one against current sources, drafts the article, generates its featured image, uploads the media and creates the WordPress post, with approval and scheduling logic standing in front of publication.
CHALLENGE
What made it difficult
A generator on its own produces text nobody wants to publish. With no research step it repeats what the model already believed, and with no tracking it will cover the same topic three times in a month.
Publication is also where an automated pipeline can do real damage. Anything that posts straight to a live site with no approval gate is one bad draft away from being switched off for good.
WHAT WE DID
The work delivered.
- 01
Topic generation and research
Scheduled topic generation aligned to selected services, with the resulting search queries researched through Tavily so a draft is built on current material.
- 02
Drafting and imagery
Structured article generation with Gemini, plus a generated featured image uploaded into the WordPress media library.
- 03
Publishing
Post creation and updates through the WordPress REST API, including excerpt, category and featured media handling.
- 04
Approval, scheduling and tracking
Approval and scheduling logic ahead of publication, with a run identifier and publishing record per article and an email notification on completion.
ARCHITECTURE
How it fits together
- Scheduled triggers driving topic generation, so the pipeline runs on a cadence rather than on request
- Research separated from drafting, which is what keeps an article grounded in something other than the model
- Structured output between generation and publishing, so the WordPress step stays deterministic
- A tracking record of runs and articles, giving the pipeline a memory of what it has already covered
TECHNOLOGY
The stack
- n8n
- Google Gemini
- Tavily
- WordPress REST API
- Google Sheets
- AI Image Generation
OUTCOME
What exists now.
An article moves from topic to scheduled WordPress post without a person carrying it between five tools
Nothing publishes without clearing the approval step, which is what keeps the pipeline trusted enough to keep running
The publishing record stops the pipeline covering the same ground twice
NEXT STEP
Tell us what you're building.
Bring us the problem with its real constraints attached. We will tell you what we would build, and what we would not.