My content system ran itself for 3 days while I was down with an arthritis flare
May 6, Tuesday. Arthritis flare. Bad one. My husband helped me get dressed. I made it to the couch and stayed there until Thursday afternoon. Friday I checked the dashboard. Three days of content had shipped without me touching anything.
The system: a Supabase marketing_queue table holding 7 to 10 days of content ahead, a publishing agent called Comet that reads from the queue on a schedule, and a Sunday scheduled task that generates and validates the week's batch.
What ran while I was down: 6 posts published across LinkedIn and Mastodon already in queue from Sunday's batch, 1 book promo cycle completed, queue status green with no stuck items.
What it cost to build: 2 weeks of early mornings and evenings in February, Supabase free tier within limits, no additional paid infrastructure for the pipeline itself.
What I would tell myself in January: build the queue before you need it. The Tuesday does not give you warning. I built in February thinking I was getting ahead of a busy spring. I was actually building for a specific bad week I could not have predicted.
Six web apps, multiple books, a full-time senior role. The system is what keeps any of it visible when I am not.
If you are running multiple projects and treating content as the thing you will get to when the workload lightens: the workload does not lighten. Build the system in the gaps you have now.
This is a strong story because the value is not “content automation.” It is continuity when the founder is unavailable. That is a much sharper angle than scheduling posts.
The queue-first system is the real insight here: content already validated, staged, and ready before life interrupts. Most tools focus on generating more content, but the pain you’re describing is different. It is keeping visibility alive across multiple projects when your energy, schedule, or health suddenly drops.
If Comet becomes more than an internal publishing agent, I’d think carefully about the brand frame. “Comet” is nice, but for a broader autonomous workflow layer that keeps projects moving in the background, a name like Xevoa.com would fit the system-level direction better.