Project · Content automation
Multi-Platform Content Automation Pipeline
One draft goes in. A finished article, a set of branded assets, and a scheduled week of platform-native posts come out — published to the blog and across Instagram, LinkedIn, and Pinterest, hands-off after a single approval.
One draft in, a whole publication out — article, assets, and a week of scheduled posts.
One draft in, a whole publication out.
Not a straight line. One approval gates the article, then it fans out into asset generators that run in parallel — all converging into a single publish queue, which ships the blog through GitHub CI/CD and then schedules the posts out to every platform. Each node is a step you can run, swap, or rerun, named by the tool it leans on.
one human approval — the only gate in the run
Claude human Write the article draft → article the article fans out — every asset below generates in parallel
Claude Pull-quotes article → 10–15 quotes 
Gemini Cover concept → image ↻ the infographic re-checks its own text and repairs until it matches
Gemini · Pro Infographic article → poster 
Claude · Python Carousel spine → 9 slides 
Excalidraw optional Diagram concept → figure 
Python · Playwright Quote cards quotes → PNG set ↑ all assets converge into a single publish queue
Claude Publish queue → publish.yaml 
Astro Build & verify assets → static site a git push triggers the CI/CD that builds & deploys the blog
GitHub git push commit → main GitHub Actions CI/CD deploy push triggers build 
Zernio Schedule posts URL → calendar ↓ then the queue fans out to every platform
Instagram carousel · cover · cards 
LinkedIn PDF · cover · cards 
Pinterest pins · infographic The parts, and what each one owns.
Five composable steps with clean inputs and outputs, wired into one orchestrated run — plus a brand-context layer (the same VOICE.md and DESIGN.md my brand-systems skills produce) so the whole thing can be pointed at a different brand.
- 01
An article step that turns a rough draft — an idea, a transcript, a pasted link — into a finished piece in a defined voice, with one human approval on the spine and headline before anything downstream runs.
- 02
An asset factory that runs in parallel off that one article: pull-quotes (which feed branded quote cards in two themes), a wordless cover, a text-accurate infographic poster on its own verify-and-repair loop, a diagram, and an 8–9 slide carousel — every branch checking its own output before they all converge into one publish queue.
- 03
A blog publish step that compiles the article and its assets into the site and verifies the build — then a git push to GitHub triggers a CI/CD pipeline that builds and deploys it to the live blog and writes the URL back into the queue.
- 04
A distribution step that routes every asset to the platform and format it fits, then schedules the posts across Instagram, LinkedIn, and Pinterest — idempotent, so a re-run never double-posts.
- 05
Brand context kept in one place — the brand's voice and design captured as a VOICE.md and DESIGN.md, plus its accounts — so the whole pipeline can be pointed at a different brand without rewriting a single step.
One human decision, then it runs itself.
The only place a person is needed is the article — approve the spine and pick the headline. After that, every asset is generated and self-checked, the blog build is verified before it ships, and the social posts are scheduled deterministically. The run is idempotent end to end, so it's safe to re-run without double-posting or duplicate work.
- Approve the article
- Generate assets in parallel
- Ship the blog via GitHub CI/CD
- Schedule across platforms
- Posted, on its own
One draft, four platforms.
The same idea arrives native to each place — long-form on the blog, a carousel and cards on social, tall pins for discovery — from the one run.
Distribution stops scaling with effort.
A single draft now becomes a full, on-brand publication scheduled across four platforms in one run — minutes of work instead of a day of copy-paste. The one decision left to a human is the article itself; everything after it — assets, build, scheduling — runs unattended and is safe to re-run.
-
Stay present without living in the apps
Keep your brand showing up on Instagram, LinkedIn, and Pinterest every week without spending your days inside any of them — you write the idea, the pipeline keeps the calendar full.
-
Get a second life out of old posts
Feed in something you published once — a past article, a talk, a link — and it comes back out as a fresh round of platform-native posts.
-
Turn a talk or a call into a week of content
A recorded webinar, a podcast episode, or a sales-call transcript goes in; a finished article and a week of scheduled posts come out.
-
Stand up a new brand's channels from day one
Point it at a new product's voice and design, and it publishes on-brand across every platform before you've hired anyone to run them.
Write the idea once. The pipeline does the other forty-nine posts.
Have a backlog of ideas going nowhere?
Tell me where you publish and how often, and I'll come back with what a pipeline like this would turn one draft into — and what it would take to stand up.