My body quit for 3 days. The pipeline did not. Real numbers from last week.

Last Tuesday my rheumatoid arthritis flared bad enough that I needed help getting dressed. I spent three days mostly on the couch. The content pipeline I have been building for six months published nine items without me touching it.

Here are the actual numbers, no polish.

Queue status as of Sunday morning:
Published all time: 44
Rejected: 22
Manual publish needed: 15
Human review required: 8
Approved, waiting: 1

The rejection number is the honest story. 22 items rejected means 22 pieces of content that failed the database validation check. Mostly content submitted as planning briefs instead of finished copy. The validator catches it. I still have to clean it up.

What the pipeline is: six web apps (ContentForge, Momentum, HomeGrown, PillPal, PawFormance, Palette Pro) plus personal content (17 books, LinkedIn, blog). The pipeline generates content for all of them, validates it, queues it by platform and week, and publishes approved items on schedule. Local Python process handles the actual publishing. Supabase is the queue.

Revenue data this week: Stripe and PostHog were not connected to this automated run. Next week will have numbers. Based on what I saw last week, ContentForge is getting active use. The Lovable apps are generating some organic signups from directory submissions. No dramatic revenue events to report.

The thing I am fixing this week: tighter pre-validation in the Sunday generation step so drafts do not enter the queue at all.

The question I keep sitting with: if the system is consistent and I am not, am I building a business or a process? I think both, but the distinction matters for what I focus on next.